SDS 319: The Path to Data Visualization

Podcast Guest: Jonathan Mucha and Ogo Ezeofor

December 4, 2019

An in-depth conversation about the details of data visualization, beyond just the technical aspects, and into the processes and psychology you need to follow to be successful for your stakeholders.

About Jonathan Mucha
Jonathan is what he likes to describe as a computer nerd and problem solving addict. All his life he’s been looking for new and interesting problems to solve, and while he’s always been happiest when he’s been able to use his computer to do so, that hasn’t stopped him from enjoying anything from rebuilding cars, designing increasingly elaborate hydroponics systems, or doing graduate level quantum mechanics. Currently he works daily as the operations manager for a hardwood flooring distribution company where he’s been playing around data science to tease out a few extra dollars in profit, and he is a proud co-founder of the online hardwood flooring sourcing platform RevelWoods.com, where he applies his problem solving abilities to do everything from help improve and debug the website, to photography, and logistics. The story of how he came to data science is a an interesting one, and he looks forward to sharing it on the podcast.
About Ogo Ezeofor
Ogo is a proven Data Scientist and Business Manager with over 7 years of experience uncovering actionable insights and guiding organizations through strategic analytics deployment. He has generated over $3 million in revenue by using advanced data analytics to solve complex problems and drive decision making for start-ups, mid-size and Fortune 500 companies.
Stakeholders want more than a chart on a page — they want data that tells a story. Ogo helps medium to large-size companies make better business decisions with thoughtful data visualization and automation techniques for live business reporting. The series of dashboards features advanced drill-down and filtering capabilities- allowing teams to uncover cost variances, increase operational efficiency and maximize revenue growth. The visually impactful dashboards, coupled with automation, helps clients spend less time compiling reports and more time analyzing insights and taking action.
Overview
Jonathan and Ogo are great examples of how data scientists, across the bord, tend to be incredibly helpful and giving people. Ogo pursued a career as a self-taught data scientist and analyst after he decided he didn’t want to pursue a career in medicine. He found himself struggling with Tableau and found our course which propelled his learning and visualization skills. Jonathan has a more traditional data science background in math and physics. A few years ago he was trying to expand his CS skills and stumbled into machine learning courses. 
The intellectual transition between visualization and machine learning is an interesting one. Jonathan prizes the ability of visualization to explain things quickly, especially to those who are not necessarily adept in data but are decision makers in their company. Ogo finds the visualization important to provide information for quick decisions to stakeholders. Ogo’s approach is interesting, he doesn’t necessarily look at the data itself until his 3rd step in visualization, first attending to the needs of the client. This puts the priority questions to the forefront rather than the data itself. Jonathan’s methods, up until hearing Ogo’s methods, has been the exact opposite.
As far as tools go, Ogo uses Tableau primarily as well as Power BI. He likes the advanced features and clean interface of Tableau. Occasionally, the client’s needs affect the tools that Ogo decides to use. About 60% of the time they choose their visualization software while 40% of the time they don’t have a tool preference. Jonathan prefers Tableau but has recently picked up using Dash. He’s been practicing using it a little more after experiencing R Shiny at DSGO and finds it’s easier to find future insights through a combination of that and Dash. What this shows is how broad a career in data visualization can be and how many different paths you can take. Personally, I love making something new and something that has never existed before—and obviously I love Tableau.  
What about tips they can offer? Both of them go to confeerences. Ogo tries to go to about two conferences a year while DSGO was Jonathan’s first conference in the hopes of making professional relationships with fellow data professionals. Besides DSGO, there is a Tableau centric conference in Las Vegas and Ogo goes to local meet ups and seminars. He recently attended Art + Data which is a visualization seminar. Ogo suggests using space effectively in visualization and striving to tell a meaningful story in the data: keep it simple and use color. He also suggests using KPIs effectively to know what you’re aiming for and, simply, make it look nice. Jonathan suggests downloading and using Kite, an ML recomendation system. He also highly suggests communication with the person using the dashboard as an important process technique and focus on the psychology of your client. My own tips include aggregation and granularity, keeping an eye on how the visualization will be consumed and get yourself certified. 
In this episode you will learn:
  • Jonathan & Ogo at DSGO Tableau workshop [6:40]
  • Jonathan and Ogo’s backgrounds [9:15]
  • Machine Learning vs. Visualization [16:50]
  • Tools [26:55]
  • Tips [42:18]
Items mentioned in this podcast:
Follow Jonathan
Follow Ogo
Episode Transcript

Podcast Transcript

Kirill Eremenko: This is episode number 319 with Data Visualization Experts, Jonathan Mucha and Ogo Ezeofor.

Kirill Eremenko: Welcome to the SuperDataScience Podcast. My name is Kirill Eremenko, Data Science Coach and Lifestyle Entrepreneur and each week we bring you inspiring people and ideas to help you build your successful career in data science. Thanks for being here today and now let’s make the complex simple.
Kirill Eremenko: This episode of the SuperDataScience podcast is brought to you by our very own Data Science Insider. The Data Science Insider is a weekly newsletter for data scientists, which is designed specifically to help you find out what has been the latest updates and what is the most important news in the space of data science, artificial intelligence and other technologies.
Kirill Eremenko: It is completely free and you can sign up at www.superdatascience.com/dsi. And the way this works is that every week there’s plenty of updates and seemingly important information coming out in the world of technology. But at the same time it is virtually impossible for a single person on a weekly basis to go through all this and find out what is actually really relevant to a career of a data scientist and what is actually very important. And that’s why our team curates the top five updates of the week, puts them into an email and sends it to you.
Kirill Eremenko: So once you sign up for the Data Science Insider, every single Friday you will receive this email in your inbox. It doesn’t spam your inbox, it just arrives and has a top five updates with brief descriptions. And that’s what I like the most about it, the descriptions. So you don’t actually even have to read every single article.
Kirill Eremenko: So our team has already read these articles for you and put the summaries into the email so you can simply just read the updates in the email and be up to speed in a matter of seconds. And if you like a certain article you can click on it and read into it further.
Kirill Eremenko: And so whether you want great ideas that can be used to boost your next project or you’re just curious about the latest news and technology, The Data Science Insider is perfect for you. So once again, you can sign up at www.www.superdatascience.com/dsi. So make sure not to miss this opportunity and sign up for the Data Science Insider today, and that way you will join the rest of our community and start receiving the most important technology updates relevant to your career already this week.
Kirill Eremenko: Welcome back to the SuperDataScience Podcast ladies and gentlemen, super pumped to have you on today’s episode today we have not one but two special guests, Jonathan and Ogo. I met Jonathan and Ogo at DataScienceGO 2019 in San Diego and we really hit it off. We were actually at the Tableau workshop that I was running and they were helping out there because they got through their visualizations faster than many people there and instead of just sitting around doing nothing and they helped out others with their visualizations.
Kirill Eremenko: It was a really cool community building thing that we had and we had fun at that workshop. And then I decided to bring them on the podcast to talk about visualization and that’s exactly what you will hear in today’s episode. We are going to be talking a lot about visualization. Specifically we’ll dive into the following things.
Kirill Eremenko: You will hear about career paths in data science and how even in the space of visualization you can structure different career paths. So depending on how you want to build your career in data science and what you’re passionate about, you can take a career path that’s similar to Ogo’s or a career path just so that similar to Jonathan’s and I think you’ll be very excited to hear the contrast between the two.
Kirill Eremenko: Also, we’ll talk about different tools that are used for visualization from Tableau to Power BI to Dash in Python to Shiny in R and some other tools will be mentioned. So I think you will find that part exciting as well. We’ll talk about the different approaches and Ogo will share his methodology, which all of us on this podcast found very interesting. So you’ll hear a really cool methodology, which you can apply in your visualizations already today.
Kirill Eremenko: Then we’ll talk about tips. We’ll share some tips. Each one of us will share three tips on visualizations, which you can use in your career to really skyrocket your skill. So make sure to hang on for that. It is quite a longer podcast, but towards the end we’ll share these tips and I think you’ll find them very insightful.
Kirill Eremenko: And finally we’ll talk about certifications, specifically about Tableau certifications. I passed my certification a few days ago from the moment when I’m recording this audio. I really enjoyed it. I’ll share my experience here and also there’ll be a few hints on spoiler alerts about what’s coming up in January and February, 2020.
Kirill Eremenko: So make sure not to miss out on that. And I can’t wait for you to check out this episode. So without further a do, I bring to you Jonathan Mucha and Ogo Ezeofor, Data Visualization Experts.
Kirill Eremenko: Welcome back to the SuperDataScience Podcast ladies and gentlemen, super excited to have you on the show. Actually double excited because we have two guests today. We have Ogo calling in from Seattle and Jonathan from New York. Guys, welcome to the show. How are you doing? 
Jonathan Mucha: Great. How are you doing Kirill?
Kirill Eremenko: Amazing. Ogo how about you? How’s Seattle these days?
Ogo Ezeofor: Excellent. Glad to be on with you.
Kirill Eremenko: That’s so cool. Well, it’s been awhile, hasn’t it? When was it, we met at DataScienceGO? That was what? End of September, right? What have you been up to since then?
Ogo Ezeofor: It’s been a fun time at DataScienceGO, I felt like I learned a lot there and it was great, teaming up with you and helping you at your-
Kirill Eremenko: Workshop.
Ogo Ezeofor: Workshop yeah.
Jonathan Mucha: That was really fun. It was great meeting you both right there. That was … I think we made a pretty cool team. The DataScienceGO since then it’s just been work work work and trying to be a dad at the same time, which is fun.
Kirill Eremenko: Nice. Very cool. So just for the sake of our listeners, I love that situation. So we had a workshop, I was running a workshop on Tableau. There was like about 60 people in the room and everybody was different levels. So you guys were obviously more advanced. There were some beginners, there were some intermediate data scientist visualization people and experts. And so how did it happen that we started working as a team. Jonathan, how would you describe that situation? How did it all … because it all happened so naturally.
Jonathan Mucha: Yeah, it did. It was pretty cool. So I think you maybe spotted the fact that I was already done with what you were telling us to do. You go, “What are you doing?” “Well, I made my dashboard” You said “Well there’s all these people here. Get up and help.” So I was like, “Okay, yeah, I’ll get up and help.” I didn’t realize that was the front row. It was the only seat left right? So I think you said, “Hey, does anybody need help? And I turned around and I’m like, I don’t know, 10 or 15 hands in the air. And I was like, “Oh, okay, cool.”
Jonathan Mucha: So I just take some people or maybe you directed me to see some people. And we sat down and started working and it was fun. I actually ended up helping two really cool people. So one was Bowen Chen, who I met actually in the SDS Slack channel and I didn’t even realize I was working with him and then later on the day I was like, “Oh wait, we worked together. We’ve worked together in the Slack channel.” So, that was fun.
Kirill Eremenko: Very cool. Awesome. Then Ogo, how did that make you feel helping others?
Ogo Ezeofor: Yeah, it made me feel great. I feel like it came kind of natural, sitting next to people who are kind of struggling, being able to just assist them on adding a parameter or just adding drill down or just completing certain tasks. It became pretty natural and I feel like I love helping people. So it was a pretty fun and exciting experience just to be out there and contributing.
Kirill Eremenko: Yeah. Like I totally loved it. It was very unexpected for me because I thought, all right, we’ll just go progress through the dashboards together. Because there was so many different levels, everybody’s progressing different stages and could have been like some people are far behind, some people are getting bored because they are far ahead, but data science is such like giving and friendly people that … I don’t know, like this just works out so well. Like they’re helping each other. Yeah. And that was an extraordinary, I mean, even for me, just to see that it was really extraordinary.
Kirill Eremenko: So you guys are really good in visualization. So maybe to get a bit of a feel for how that happens, could you give us a bit about, like a big … a quick rundown on your background? So Ogo let’s start with you. Like when did you get into visualization and how come you know Tableau already?
Ogo Ezeofor: So I’m a bit of a nontraditional, self taught analyst. So I was born and raised in on the North side of Chicago. The go, Cubs. My parents, they immigrated from Nigeria about 35 years ago. And in our house they pretty much promoted us to … they wanted us to be either a doctor or an engineer. Those are the two career paths, right? So essentially I stumbled into data science after medical school didn’t work out, actually wanted to become like an orthopedic surgeon. And until I realized I didn’t want to spend the rest of my life at a hospital.
Ogo Ezeofor: So, essentially I began my data science career about five years ago when I was able to quickly propel myself in the field by becoming a business intelligence analyst for a small healthcare in a healthcare consulting company. And it’s a [inaudible 00:10:29] Washington that’s slightly outside of Seattle.
Ogo Ezeofor: And essentially what I did was I came into work and I was banging my head because I just could not figure out Tableau. I was just having trouble kind of understanding how to build these visuals in a meaningful way. And so I did that for a couple months and I just was not making any traction.
Ogo Ezeofor: So I decided, one day I decided that I would, I need to take a course on this, I need to figure out how to become more efficient and Tableau. So I found your course, Kirill and surprisingly it helped me-
Kirill Eremenko: Awesome.
Ogo Ezeofor: Dramatically improve my Tableau skills. So from there I just started taking more courses online. More like Udacity, Udemy, just taking as many courses on data science as I could. And that really propelled me and my career and helped me become the data scientist that I am today. So, that was kind of the initial start of how I got into data science and then from there I was able to work for Unilever and Microsoft as a data scientist.
Kirill Eremenko: That’s so cool.
Ogo Ezeofor: So, the experience that I gained and the knowledge that I kind of absorbed from your courses as well as from others, kind of put me in a position where I was able to showcase my data science skills on a high level too.
Kirill Eremenko: That’s so cool. That’s very, exciting. So within five years, in those five years, you already worked for Microsoft and Unilever, just by learning stuff online. That’s really cool. I know you’re running your own freelance operation in visualization right?
Ogo Ezeofor: Yep. Definitely. So, yeah I was able to generate over $200,000 in less than a year.
Kirill Eremenko: Wow. No way.
Ogo Ezeofor: As a freelance data consultant.
Kirill Eremenko: That is so cool. Congrats man. That is awesome. 200k in a year as a freelancer. Wow.
Ogo Ezeofor: So yeah if I can do it, you could do it too right?.
Kirill Eremenko: That’s awesome. That’s really cool. Congrats.
Ogo Ezeofor: Thank you.
Kirill Eremenko: And Jonathan, how about you? Your story is a bit different. Tell us about you.
Jonathan Mucha: Yeah, so my story is quite a bit different, but my background … undergrad background was in physics and math, right? So [inaudible 00:12:44] physics and [inaudible 00:12:45] math. And so data analysis… For me all started with Excel back in 2003, and so I did that for whatever that was my undergrad. And so I’ve always loved solving problems with my computer. That’s that’s what I do. I just like doing code, I see a problem where I see an opportunity, I’m like, “Oh, I bet I could write something that would make this go smoother.” Right?
Jonathan Mucha: So in about 2017 I was trying to expand a little bit of my CS knowledge and stumbled onto I think Andrew Ng’s course on Coursera that was machine learning. And I had sort of heard machine learning as a buzz word and people talk about it and I thought, “Well this is probably something to do computer science, let me look into this.”
Jonathan Mucha: So I started doing it and I was like, “Oh, this is kind of fun. There’s an awful lot of MATLAB, which I hadn’t used in a long time,” but it was fun. And so when I finished that course I thought, “Well let me look into this a little bit.” And I actually, I started looking for podcasts for data science, right? So I stumbled onto a bunch of them, found a bunch and yours was the most interesting to listen to.
Kirill Eremenko: Wow. And now you are on the podcast.
Jonathan Mucha: I know. How much fun is that, right? So, I was listening to your podcast and I thought, “Huh, this guy seems like he has some cool ideas. I’ll check out his website.” I signed up for your website, took the Tableau course and I don’t know a bunch of the other ones too. And that’s sort of got me started probably early 2018 all the way through 2018 and up until early 2019 and my son was born early in February 2019 so when I stayed home with him for a couple of weeks afterwards while he napped, I stayed up and learned as much data science as I could in that time.
Kirill Eremenko: Amazing.
Jonathan Mucha: So that’s sort of how I got started. And since then, I mean I’m currently a manager of a hardwood distribution company, hardwood flooring distribution company and also have a startup that I’m co-founder of, that’s a Hardwood flooring sourcing platform and- 
Kirill Eremenko: Very different to data science. You couldn’t be more different.
Jonathan Mucha: Very different from data science. Yeah, it’s very physical and labor intensive. But what I’ve been able to do is think through the stuff that I’ve learned from your courses, other courses, reading online and take those ideas and say, “Hey, I can apply data science here. Even though they’re, kind of look like they’re totally opposing things. There’s room here to apply data science and make some progress.”
Jonathan Mucha: And I implemented something earlier this year that was able to knock down our inventory value. Right? So we carry about a million dollars worth of stuff in our warehouse and I was able to get that down about 15% to 20%, hovering around 800,000 instead of a million, which is great because that freeze up…
Kirill Eremenko: That’s a lot. That’s like what Ogo made in a year. You saved your company as well in a year.
Jonathan Mucha: Yeah. And what was really remarkable was that it didn’t take very long, built a system, implemented it and I’d say within two months there was market difference in what was happening in our business. So it’s pretty wild.
Kirill Eremenko: Wow. That is really cool. I like how you mentioned you … both of you guys have different ways you started. Ogo you needed to get your head around Tableau. You found the courses, you took the courses, Jonathan, you started with machine learning, then you listened to some podcasts. Then you did the visualization courses.
Kirill Eremenko: So what I was curious about is this transition of Jonathan that you went through from machine learning to visualization because data science, you can do data science as machine learning, you can do data science as visualization, data preparation, deep learning, storytelling. There’s many areas in data science.
Kirill Eremenko: So what are your thoughts both of you guys, what are your thoughts on machine learning versus visualization? Where is data science for you?
Jonathan Mucha: So that’s a great question because it’s all part of data science, right? And machine learning is great and I like it and I think it’s important it’s useful. But when it comes down to your boss saying, “Well, why are you spending this much time doing X?” Whatever that thing is, right? You need to be able to say, “Well look, here is why I’m doing it and this is what’s happening.” And so if you can do that with a compelling visualization, then it’s a lot easier than trying to say, “Well, if you take the average of this thing and then you add the standard deviation, you end up with this whatever” and their eyes have glazed over. My eyes are glazed over now just saying that.
Jonathan Mucha: But if you show them a graph and say, “Look, this is where sales have moved from. We used to do a lot of sales here in this category. Now we’re doing a lot of sales here in this category.” That’s two seconds. And they understand it perfectly. So visualization is a super powerful aspect of data science when it comes to communicating the results, which is really what it’s all about. Right?
Kirill Eremenko: Got you.
Ogo Ezeofor: I completely agree. A lot of times as data scientists, we tend to focus about 80% of our time doing data prep work and transforming data, but, and then reserving 20% of the time for visualization. But, and if you really think about it, the visualization aspect is very critical and essential, especially in the business world. Like Jonathan mentioned, that’s the client facing product. So everything that we’re doing behind the scenes is very important, but also we have to spend a lot of time understanding how we want to convey that message. Once that data goes into some kind of visualization and being able to use color, for instance, to quickly identify areas of opportunity so that the stakeholder can make quick decisions, quick and informed decisions is very essential in the data science community. 
Kirill Eremenko: Got you. Totally agree as well. So totally agree. And Ogo I was actually curious, to your point that, that’s very important for businesses. What kind of projects have you lately seen in your freelance startup? What are businesses, the needs of businesses where you help them on? What do they need in terms of visualization, is there anything you can share with us?
Ogo Ezeofor: Yeah, definitely. So, a lot of times, some companies that I work with, they kind of just put a bunch of worksheets together on a page and sometimes it can get kind of confusing and kind of hard to follow. It’s not necessarily telling a story. So what I like to do is I like to take a step back before the visualization and I have a three step process that I kind of rely on.
Ogo Ezeofor: And the first step is the discovery step. And in that step the goal here is to build the requirements quad. So the requirements quad is essentially you’re taking four different elements of kind of gathering requirements. 
Ogo Ezeofor: So like let’s say the first quad will be you’re identifying goals and objectives of the dashboard. So what would this company like to achieve through this particular dashboard? Right? What are some of the goals and the challenges, right?
Ogo Ezeofor: And then the second quad we’re looking at the audience and the user. So who will be using this dashboard on a daily, weekly and monthly basis? So would that be like the CEO or the chief, let’s say the marketing officer or director?
Ogo Ezeofor: And then the third quad are some priority questions. So what are the top questions that the business would like to have answer, such as what is my profit trend or what is my customer distribution? Or what is the product distribution, right?
Ogo Ezeofor: And then the fourth quad is some other requirements. Does this visualization, does it have to be built in a specific BI tool such as Tableau or Power BI? Would the user like to have drill down capabilities? Right? So after this stage of discovering those requirements, I tend to move on to the next stage, which is the drawing stage.
Ogo Ezeofor: And here I like to take what I gathered from the first step and then build out a Mockup dashboard, right? I will follow with everything that we’ve learned, in the requirements stage. So here we create a rough draft of a dashboard and we place all the KPIs and the bar charts in a meaningful way so that it’s telling a story.
Ogo Ezeofor: And then after doing this, I like to get some client feedback in terms of, is this, kind of in-line with what they are looking to achieve or accomplish with their dashboard build? And if they are in agreement with that, then I move onto the third step, which is the development step.
Ogo Ezeofor: And that’s when we dive into to the data. So we try to figure out, okay, what data sources are needed in order to build out these KPIs and these dashboards and these worksheets, sorry, on these dashboards. So essentially figuring out whether we need marketing data, do we need, data from the website. So let’s say e-commerce data or data on like Google analytics for instance.
Ogo Ezeofor: So after kind of developing the dashboard in Power BI, we’re then ready to present that to the client. So it’s, this three step process that makes it so that we cover all the bases and we know exactly what the end user is … what’s most valuable for the end user and how to display that data in a meaningful way.
Kirill Eremenko: Wow. That’s a really cool process of what I really like about it is first of all it’s very structured and second that … So you don’t actually look at the data available to you. If I’m understanding this correctly, you don’t look at data available to you until step three. So you first identify all the requirements, needs, who’s going to be using it, the priority questions, what tool you’re going to be using for this? Then you draw up a blueprint you don’t even know what data you have in the organization at this stage. You’re just attending to the needs of the clients. 
Kirill Eremenko: Then you run it by them, you get their feedback and then only in stage three do you look for the data? Is it, there like, how can I get it? How can I source it? And so on. Is that the correct summary?
Ogo Ezeofor: Yeah, definitely. And I think that approach is definitely very valuable because instead of … without looking at the data or let’s say for instance if a company has access to let’s say several data sources, it can get very overwhelming as a data analyst or data scientist to just come into an organization and then start trying to kind of throw worksheets onto a dashboard without having a clear understanding of what the client is trying to accomplish with the dashboard. So I think having that step wise approach add some additional clarity that is needed so that you’re only focused on the data that is required to answer those specific priority questions.
Kirill Eremenko: Interesting. Have you ever come across a situation where it’s opposite where you do step one and step two and you’ve drawn everything up and then you go to look for the data and the data’s not there? It’s not that they have too much data that it would overwhelm you. Like they don’t have enough data for you to prepare this dashboard that you’ve planned out.
Ogo Ezeofor: Yeah, actually, surprisingly, I’ve never actually had that problem. It seems like there’s always enough data to answer the given question.
Kirill Eremenko: That’s cool.
Ogo Ezeofor: But yeah, I think in a situation like that, it would just be, that would kind of generate another question like, okay, so how can we get the data that will help answer that particular question.
Kirill Eremenko: That’s very cool. That’s very cool. And in deed if the question is answering, the data shouldn’t be a constraint, the organization should find a way to get the data. Very cool thanks for sharing. And Jonathan, what do you think of Ogo’s methodology, do you have a methodology of your own? How do you develop dashboards?
Jonathan Mucha: All right, so Ogo’s methodology just became my methodology.
Kirill Eremenko: I was thinking the same thing. When Ogo was talking I was like “Okay, that’s how I’m going to approach dashboards from now on.” That’s awesome.
Jonathan Mucha: That’s beautiful. So I have a little project that I’m working on now that I’m going to try and apply this methodology too and just see how it works out because it’s beautiful. I love the idea of not even bothering to consider data until you’re done with the design and say this is what I need because to Ogo’s point, right? It can get overwhelming really fast. Well, I’ve got this data, I’ve got this data, I’ve got this other data. So to just eliminate that and say, this is what I need, this is what the client needs. Okay, now I’ll go in cherry pick the things that I need to go and get as far as data goes. That’s beautiful.
Jonathan Mucha: So up until now, my methodology has been sort of the opposite of what Ogo just said and I realize now that I’ve been totally screwing it up.
Kirill Eremenko: Yeah, that’s why, Ogo makes the big bucks.
Jonathan Mucha: Yeah. Clearly.
Kirill Eremenko: He’s got a down path.
Jonathan Mucha: I liked it. That was beautiful.
Kirill Eremenko: Awesome. And Ogo, let’s talk about tools for a bit. Ogo, you mentioned Tableau or Power BI, there’s quite a few tools out there for visualization. Tableau, Power BI, QlikView, Qlik Sense, probably a couple of others not as dominant but still exist out there. Well, what are your thoughts guys? Like what, what’s the best one? What’s the second best? What’s your favorite? And so on. I don’t know, let’s maybe start with you Ogo since you brought that up.
Ogo Ezeofor: Yeah, I mean, my primary analytics tool that I use in terms of visualization is Tableau, followed by Power BI and Tableau, I just … I’ve had more experience in Tableau in a little over five years and I found that Tableau has very easy user interface and it seems like they’ve had a long time for research and development and figuring out what the end user would like to accomplish and how to make certain elements of dashboard design easier.
Ogo Ezeofor: And so, it seems like Power BI, it’s still catching up, but it seems they’ve made very large strides and they are … it’s a fairly nice tool as well. I see value in both, but majority of my projects have been in Tableau and some of them more recently have been starting to use Power BI and generally companies that are looking to use Power BI now one of the key things that they see as a benefit is the fact that it’s integrated with the Office 365 suite and it’s just more seamless in terms of, compatibility.
Ogo Ezeofor: But Tableau is definitely has some advanced features and a very clean user interface and you can kind of complete certain tasks with the click of a button versus Power BI sometimes it takes a couple of clicks, but I think they’re working on that and I think in time Power BI will definitely catch up with Tableau.
Kirill Eremenko: So does the choice of tool depend on the client or is it sometimes you go in and you’re like, “Oh, this tool would be better for this job or this tool will be better for this job?
Ogo Ezeofor: I would say sometimes it depends on the client because they are pretty adamant sometimes about using Power BI. Like if they have a full integration of like Office product, but sometimes they ask for recommendation, like what tool do you recommend that we use and then that’s kind of where I kind of try to gauge their situation and get a recommendation based on their current infrastructure.
Kirill Eremenko: What’s the percentage or in what cases, what’s the percentage of cases where you have to use the tool they have and what percentage of cases do they ask for recommendations?
Ogo Ezeofor: I would say it’s about 60% to 40% so 60% of the time they have their choice of a visualization software that they are needing help with. And I’ll say the other 40% of the time it’s like they don’t know what tool to use. They just have a bunch of Excel spreadsheets and then they’re trying to find a streamline the process to build visualizations. So then they rely on recommendations. 
Kirill Eremenko: Okay. And so out of those 60%, what percentage would you say already use Tableau and what percentage already used Power BI from your experience?
Ogo Ezeofor: I would say a majority. Oh actually no. I’ll say like I’d say like 70% of the clients are using Tableau, but 30% are using Power BI. And actually what I’m starting to see is that more corporate clients are using Power BI, whereas some of the smaller and mid sized companies, rely on Tableau. And I think it has something to do with at a corporate level, they’re pretty dependent on Microsoft products and they want to stay in line with that. So they, they tend to focus on implementing Power BI so that there’s compatibility there. 
Kirill Eremenko: So, that’s very interesting. I guess it kind of makes sense in terms of their corporate … what’s it called? When Microsoft is dominant player with a lot of these corporates and … but basically what I was getting to is that it makes sense to know both to learn Tableau and Power BI because that will at some point help you in your career whether you’re a freelance or whether you’re looking for jobs because always, not only will you get additional knowledge by knowing both, you’ll get insights that maybe you don’t see through Tableau, Power BI will focus your attention to something better. And Tableau as well will help you learn something in another way. But also just for your career as well, you’ll have the flexibility to go with either way. Would you, agree with that? 
Ogo Ezeofor: Yeah, I agree with that. I think it’s definitely helpful to know both for sure.
Kirill Eremenko: Cool. And Jonathan, what about you? What are your favorite tools?
Jonathan Mucha: So I like Tableau, I’ve never actually worked with Power BI, is I didn’t get through that course, but the other thing that I’ve been using recently, since actually I attended Matt Dancho’s seminar after your seminar at DSGO right?
Kirill Eremenko: Yeah.
Jonathan Mucha: And he works with … he was working with Shiny and R. And so we built the dashboards on the fly there, which was pretty fun. And I thought “This is actually really cool.” So I went through and looked for an equivalent in Python, which I guess is Dash.
Kirill Eremenko: Dash yeah.
Jonathan Mucha: So I re-implemented Matt’s code in Dash and just as a way to learn it. And so that actually I have been using a little bit more lately because I think it’s a little easier to add future predictions and insights from machine learning algorithms, right? So if I run XG boost on something and I get some output, it’s a little bit easier for me to put that up in something that is easily hostable. Right?
Jonathan Mucha: So Dash builds these little flask powered websites and you can throw it up on Heroku or something like that and have access to it right away. So just because I think it’s a little bit more fun and, and easily sharable. Tableau is expensive product if you want to be able to download things and share it.
Kirill Eremenko: That’s a very interesting comment because when I was mentioning the tool before, like Tableau, Power BI, QlikView, Qlik Sense and so on, those are all drag and drop tools. Those are all tools where you can learn it very quickly. It’s very easy to use, very intuitive and so on. But indeed there’s this a whole suite of other tools that are related to the programming languages such as the ones you mentioned, Shiny for R programming, Dash for Python, which are free. You don’t have to pay anything for them. And they’re very different. They’re not drag and drop, you have to code them.
Kirill Eremenko: So I’d love to get a bit more of your thoughts in that Jonathan, how is your … how do you feel about using either or how has your experience using either, because Tableau you just drag and drop. You don’t need to code pretty much anything except for if you have like a calculated field or something like that. But in Dash or in Shiny you have to actually code the whole visualization. You have to write it in code. So what are the pros and cons of either approach?
Jonathan Mucha: So for me Tableau is definitely something where if I need to bang something out really fast, Tableau is where to go, right? If, I’m looking for something that is going to display data over a long time period or something that’s going to run continually where I want to be able to log into a website and say how am I doing this month or what are my predictions for next month? Where should I be? What’s my sales, what are my sales going to be and have a lot more options and a way to share that with other people, I like using something like Dash. And it is a little bit more customizable I think, because you’re integrating more than just visualization, right?
Jonathan Mucha: So you can do some cool that you do the clustering thing in Tableau or the trendline thing, which is … those are great. But if I want to, use an ARIMA model and predicts next month’s sales forecast, it’s a little bit easier to work that into a Dash outline or a Dash workflow in my opinion, than it is to work it in Tableau.
Kirill Eremenko: Fantastic. I guess you’d comes that’s a like there’s a trade off, right? Like on one hand, ease of use. But on the other hand, adjust adaptability, adjust stability or versatility of the tool, right? So like as you said, you maybe can do a bit more things in Dash, but that comes at a price of, Oh, you have to code them in.
Kirill Eremenko: On another hand, Tableau, maybe you can’t do absolutely everything that you want to imagine under the sun, but at the same time, it’s much faster and easier to get traction. And I really like how we have two different, slightly different career paths if I may say in that way on this podcast, like on one hand Ogo you’re like as I understand like focus heavily on these drag and drop tools, Tableau, Power BI.
Kirill Eremenko: On the other hand Jonathan you into Tableau visualization, but also you’ve gone into a slightly different direction with the Dash as you are taking your career in other way. I think for listeners of the podcast this can be very valuable because data science is so broad, it has so many ways you can take your career. Even in the space of visualization, you can go in Ogo’s direction and focus just on or like predominantly on these drag and drop tools, and getting really good at them and bring value to customers and develop your own methodology around that. 
Kirill Eremenko: Or you can go and combine your visualization as an add-on to your machine learning skills and use something like Dash in Python or Shiny in R and still get the value from visualization but also be able to integrate machine learning models and so on. So you can choose either and depending on and what you like and what you’re excited about.
Kirill Eremenko: So that excites me personally about data science. It’s so anybody can have a career in data science that they’re excited about.
Jonathan Mucha: Yeah, absolutely.
Ogo Ezeofor: Definitely. So, Kirill what are some of the tools that you like to use in terms of data visualization?
Kirill Eremenko: Actually this is interesting because that was probably the next thing I want to talk about. I personally love Tableau. I think we all agree on this podcast. Tableau is really cool. And actually last weekend on Saturday, so like six days ago, I passed my Tableau certification, my first Tableau certification, I did the Tableau Desktop Specialist exam. I was very nervous and it was really, I was like, “Oh, this is my first Tableau exam. How’s it going to go?” But it went very well. It was really fun. It was a very nice proctor, a guy like somebody like monitoring the whole process and did all the multiple choice like 30 questions I think in an hour. Got my results, passed the test and it’s like very excited and had a lot of fun.
Kirill Eremenko: So really enjoyed that. And now I’m looking forward to my Certified Associate … Desktop Certified Associate and then I’m going to do the test of Certified Professional exam as well. Have any of you guys done the exams?
Jonathan Mucha: Absolutely not. Where do you even … where do you find them?
Kirill Eremenko: You go to tableau.com/learn/certification. They’re like 100 bucks per exam.
Jonathan Mucha: That’s awesome.
Ogo Ezeofor: Yeah. I mean I actually, I’ve always thought about completing those. I just never really got around to it, but definitely going to look into that now. Now that you mentioned that.
Kirill Eremenko: You know what guys, I got some good news for you. The reason why I’m doing these exams because, I was so inspired, because I run SuperDataScience, and the podcast and so on. So I don’t get that many opportunities to actually work on data science projects or to create visualization. Sometimes when I can, that’s really exciting. But I was very inspired by the workshop that we did in DataScienceGO. So, and that got me like, “Oh Tableau, this is so cool. I want to get into it again.” Like get even more, do more stuff with Tableau. And, then I thought, “Hey, I haven’t done the certifications.” And I talked to some people, and they told me, “Hey you should totally do the certifications.”
Kirill Eremenko: And moreover what I was thinking is like, I’ll create a course on how to do these certifications to prepare people for this certifications. So that’s why I’m doing the certifications. I did the Tableau Desktop Specialist. I’m going to do Certified Associate in a few weeks and then Professional and then probably sometime around like January, February, 2020 we’re going to launch a course on how to prepare for certified … well for all three exams for the Specialist, the Associate and the Professionals.
Kirill Eremenko: So if you haven’t taken exams by then, then I highly, welcome you to take the courses that I’m going to create and hopefully they’ll prepare you for the exams.
Jonathan Mucha: Oh, that’s awesome.
Ogo Ezeofor: Yeah, that is.
Kirill Eremenko: That’s good.
Ogo Ezeofor: Looking forward to those courses though. For sure. I will. I’ll be probably one of the first to sign up.
Kirill Eremenko: I love-.
Jonathan Mucha: God bless you.
Kirill Eremenko: That’d be cool because I love creating them because you can … it’s, similar to like in visualization you allow yourself to be creative and make, I don’t know, make something that never existed before. Same thing in courses I like to create these datasets and these challenges and so on. You guys probably remember it from the Tableau courses that they’re like, there is some pretty cool things that Tableau can do. And if you have the right dataset, the right challenge, it’s really exciting sometimes.
Ogo Ezeofor: Yeah, for sure. And one thing I really like about your courses Kirill exam, is the fact that you use real world challenges or scenarios, like in one course I think it was the Tableau A to Z or the advanced one. But you kind of like we’re using oil rigs or like oil refineries to see, to kind of see what the capacity for the oil rigs were. Right.
Ogo Ezeofor: So this is like a real world scenario and being able to then apply the data science techniques and visualization techniques to then visualize how these rigs are performing compare it to each other. Right? So, that’s the one thing that I really like about some of your courses.
Kirill Eremenko: Thanks a lot Ogo. I really believe in making, applying to real world examples and real world case studies so that, that’s for sure the case. I actually wanted to switch gears a little bit and talk about tips. So we all, we’ve mentioned visualizations tools or you’ve shared your fantastic methodology with us, which we all are going to adopt from here on down. You should really publish it like Cheat Sheet or something like that. People can download it like the Ogo visualization method, OVM.
Ogo Ezeofor: I got inspired from another company, so I have to give them a little bit of credit as well. So I think it’s decisive data. So, I went to a few conferences this year and I just try to get some little tidbits from each conference to see how, I could improve my skills as well as help others. So, I get inspiration from different people. But interesting.
Kirill Eremenko: That’s very cool. So you attend a couple of conferences. Jonathan, how many conferences did you attend in 2019?
Jonathan Mucha: DSGO was the first one I went to.
Kirill Eremenko: Okay, cool. So what made you go? Very interesting, I’d love to get both of your perspectives. Jonathan, why’d you go to DataScienceGO as your first event?
Jonathan Mucha: I think I started listening to your podcast about a week or two after DSGO18 so everybody was still pretty high on it. And I was like, “Wow, that sounded like a lot of fun.” And so as it came up this year I thought, “Well, this is a really interesting field and I’d really like to do more in it. I should probably make some friends in the data science community.”
Kirill Eremenko: Did you?
Jonathan Mucha: So the DSGO sounded fun. Yeah. I mean, Ogo and I hung out like the entire DSGO. And I met several other people that were great. So, that was really fun. Actually I met Ayodele Odubela, which is really fun and I think I’m going to be doing a mentorship with her. 
Kirill Eremenko: Nice. That’s really cool.
Jonathan Mucha: That’s going to be really fun. So yes, that was the purpose of coming to DataScienceGO this year and I think it’s filled its purpose quite nicely. I had a lot of fun, met a lot of people, and here we are chatting. So what else can you ask for?
Kirill Eremenko: Nice. Very cool. And Ogo for you. It was slightly different you went to a few conferences. So why do you go to the conferences and why did you go to DataScienceGO?
Ogo Ezeofor: So I feel like one of the most important things that you could … I guess one of the most, one of the best ways of spending money I guess as a freelancer or just as in general I think is increasing your knowledge and learning from others. So I feel conferences are a great place to learn new things and learn new methods on how to approach problems. So, that’s kind of one of the main reasons I go and I try to go to at least one or two conferences a year just to, kind of stay current and also to see if there’s something that I could apply to my day-to- day kind of operations. 
Ogo Ezeofor: So that’s kind of the main reasons why I go and then also the network. Right? And so that’s that. And then one of the reasons why I decided DataScienceGO is primarily because of you of course right? Kirill.
Kirill Eremenko: Thank you.
Ogo Ezeofor: Because I feel like you have a pretty great grasp when it comes data visualization and just overall data science. So, I was pretty impressed with what you’ve built so far. So I really wanted to check that out in person, see what that was like.
Kirill Eremenko: Awesome. Thanks. What are some of the other conferences you can recommend to our listeners? Like of course I’m biased DataScienceGO, but there’s other great conferences out there.
Ogo Ezeofor: Yeah. So there’s a conference, like Tableau conference actually, I think this weekend, I believe in Las Vegas. So that’s the conference that I would like to go to, but I’m not sure if I’ll be attending that one this year. But those are some of the type of conferences that I like to go to.
Ogo Ezeofor: And then, locally I do some, I go to some kind of meet-and-greet and some seminars locally. And one of the ones that I had recently attended was Art + Data in Seattle. So I’m, just learning how to apply certain visual techniques to dashboards so that you could help stakeholders make quick and informed decisions using Color and Flow and things of that nature. So I like to attend Meetups and seminars as well locally.
Kirill Eremenko: That’s very cool. I think Meetups are very important, yeah, as you said, meet people, network and so on. Well thanks guys for the shout outs for DataScienceGO, I’m glad you enjoyed the event. We’re going to grow it doing more next year. A fourth one and conferences I do believe are important to meet people, network.
Kirill Eremenko: Lets, switch a bit and talk about some tips, right? So we’ve talked about visualization, we mentioned some tools we love and so on. What are some tips you can share? Just quick hacks that you’ve learned using Tableau or maybe other tools that our end users can take away and after this podcast right away start applying in their careers. All right. Who wants to start a random tip that comes to your mind?
Ogo Ezeofor: I would say one of the most important tips that I can give is just try to guess build a … try to build a dashboard that tells a meaningful story and try not to overcomplicate things. So like keeping it simple is very important and using white space or just not cluttering the dashboard with, just several worksheets and charts is something that I found is a little bit more helpful. Just being able to tell a meaningful story, use color to quickly identify kind of areas of opportunity. And just not over clattering the dashboard.
Kirill Eremenko: Okay. Got you. So keep it simple and use color. First one from Ogo. We’re going to … let’s do it. Let’s make it again. Let’s do three each, right? So each one of us will do three. Ogo as good as this one. Keep it simple and use Color. Jonathan, what have you got for your first one?
Jonathan Mucha: So in terms of, if you’re going to be doing any programming, like we were talking about Dash, download and start using the program kite. If you’re going to be working in Python, it is awesome.
Kirill Eremenko: What does it do, tell us?
Jonathan Mucha: It’s a recommendation system. It’s an ML recommendation system for code completion, documentation search, and it integrates into whatever IDE you’re using and as you’re typing it’ll bring up documentation for things. So you start typing in matplot and it says, okay, this is how everybody in the internet or GitHub, this is how they’re using matplot. Like usually the next thing they type is pieplot. Do you want to do that? Yep. Cool. Okay, cool. Do you have a question about the plot function? It brings up examples and it does this all automatically so you’re not bouncing back and forth between Google “Oh wait, I got to pull up the matplot documentation because I forgot what order things go in or whatever.”
Jonathan Mucha: So I downloaded it this week and I was like “Where has this been my entire life, it’s the best-
Kirill Eremenko: That’s awesome. So it’s K-I-T-E, right? For python?
Jonathan Mucha: Yeah.
Kirill Eremenko: And it’s a separate program you install on your computer, is that right?
Jonathan Mucha: Correct.
Kirill Eremenko: And does it work with all Python tools, from Spider to other IDs?
Jonathan Mucha: So it didn’t have support for Spider but it had support for Sublime and PyCharm and Adam and Zem even if you into that kind of thing.
Kirill Eremenko: What about Jupiter?
Jonathan Mucha: No, it didn’t say Jupiter but they should definitely add Jupiter.
Kirill Eremenko: Maybe they will with time. Do you know a similar one for R?
Jonathan Mucha: I don’t, I just stumbled onto this one this week and I thought holy crap, this is great.
Kirill Eremenko: And you don’t have to just use it with visualization. You can use it for other, machine learning and Python for instance.
Jonathan Mucha: It’s for any Python program whatsoever. So, if you’re going to be … I don’t care if you’re adding two plus two or if you’re making the world’s most complicated Dash dashboard.
Kirill Eremenko: Wow. Fantastic. So my number one tip will be, I really love the way Tableau has integrated the concepts of aggregation and granularity and that is how granular is your chart? How many points are you showing? How in detail are you showing this specific visualization or this specific view or versus how aggregated is it? How are you aggregating these things? You might have chart showing results, survey results or statistical results per city or of a continent or per country or a continent, or it can show per continent. Things like that.
Kirill Eremenko: So aggregation and granularity. My number one tip, or my tip number one would be to look into these concepts, whether or not you’re using Tableau, Tableau just doesn’t really, well, I think that’s the foundational principles of Tableau aggregation granularity. That’s how the tool is built. It’s a great place to learn it.
Kirill Eremenko: But even if you’re using other tools, understand those concepts really well because that’ll help you better structure, decide and design your visualizations. When you have those two things in mind. It’s always a trade off between aggregation and granularity and finding that golden balance. This is what’s going to help you build the perfect dashboard.
Kirill Eremenko: Number two from Ogo please, sir.
Ogo Ezeofor: Yeah, so, I think it’s along the lines of what you mentioned, but I feel like there’s a way to kind of make things less complicated in a sense of you can … instead of showing a chart with a bunch of lines on it, you could say, you could build a KPI. So kind of using KPIs effectively. So let’s say you build a KPI and it’s looking at let’s say the percent, let’s say the percent change of let’s say profit for instance, over the past six months, right? So that you could say, let’s say profit has increased 30% over the past six months, right? And just a single counter or KPI. And that just quickly gives the user an understanding of whether profit is improving or declining just from looking at one KPI. So being able to utilize counters or like-
Kirill Eremenko: Numeric.
Ogo Ezeofor: KPI cards. Yeah, numeric numbers in a very concise way so that you could quickly kind of portray what’s going on with the underlying data.
Kirill Eremenko: Interesting. I love that tip because it’s very counterintuitive. You’re like, once you started doing a building visualization, you think that everything should be a picture. Everything should be an image. Everything should be a chart or a bar chart or whatever. A pie. Like it should have an element of like visual to it, but what you’re saying is sometimes don’t get carried away. Just put a number in there. Don’t forget that you also have that power to just add a KPI or a numeric representation of a single thing.
Kirill Eremenko: Sometimes you don’t need to overdo it and attach an imagery to every single or visualization type of thing to every single insight. Sometimes you just want to have a single KPI that will help inform the rest of visualizations, right?
Ogo Ezeofor: Correct. Yes.
Kirill Eremenko: Awesome. Great tip. Thank you. All right, Jonathan, your tip number two.
Jonathan Mucha: So I would say less about individual dashboards or individual visualizations as a process technique, communicating with the person that’s going to be using that dashboard constantly so that you’re constantly iterating on the idea. Like, Ogo said, this, this person needs a particular feature. And so your understanding of that feature might be different to their understanding of the feature.
Jonathan Mucha: So you build it, you have to show that to them iteratively, right? You have to say, okay, here’s what I’m thinking. Yes, no, they say no. Okay, here’s now what I’m thinking again, based on our other conversation, you have to do those things as much as you possibly can because you’ll end up wasting so much more time going down a rabbit hole that you’re going to have to backtrack if you’re not in constant communication with your client.
Kirill Eremenko: Well, wonderful. So it’s like apply an Agile kind of methodology to, or an Agile process to your development. Don’t just develop the whole thing from start to finish and then get feedback, but develop it a little bit, get feedback, do it a little bit more and get feedback and so on so that you’re always, you know, that you’re on the right track. Correct?
Jonathan Mucha: Yeah. That’s exactly right. I did steal that idea from Agile.
Kirill Eremenko: Yeah. Well that’s the best way to do it. Especially in something like visualization.
Jonathan Mucha: I agree.
Kirill Eremenko: Okay, cool. My tip number two is about presentation and how you … basically how you present. So you need to understand when you’re creating a visualization, you need to understand, is this going to be presented? Is somebody going to be consuming this without you or with you? Very two different situations. If somebody is going to be consuming without you, you have to put in all the comments, annotations, supporting text, descriptions and so on so that people are able to consume it.
Kirill Eremenko: Whether if it’s just like a storyline so that they can just follow along and understand what’s was going on or if it’s an interactive visualization you need to add documentation to it or some kind of additional texts that will help people understand, which buttons to click and so on. That’s the case number one.
Kirill Eremenko: Case number two is if they’re consuming it in your presence, if you’re going to be actually presenting this, whether it’s a story language or interactive dashboard, doesn’t matter. A PowerPoint presentation. If you copy your visualization to PowerPoint and oso n, like if you’re there, absolutely different story. My advice, if you’re presenting minimal text, you are there to present, so the attention must be 80% on you, 20% on the visualization. You are there to present, the visualization is there to support you. It’s a tool that’s assisting you.
Kirill Eremenko: The eyes and the ears should be on you, not on the visualization. You should click a slide, open a new slide or progress through visualization. People should look at it quickly and then back. Attention should be back on you, because that’s how they consume it.
Kirill Eremenko: If you put a lot of texts and a lot of pictures, images and so on onto visualization. So people are looking through it, they’re not going to listen to you. They absorb much more information through their eyes and through their ears. And therefore as they’re looking through the visualization, they’re absorbing all this information and everything you’re saying is going completely over the top of their heads.
Kirill Eremenko: So if you’re presenting, make sure that the visualization is very, like Ogo says simple and it’s designed to support you, not the other way around. You want to be the star of the show, not the visualization. So take into account, which situation are you in, is the information consumed without you or with you and make your visualization, tailor it to that specific situation. Would you guys agree?
Jonathan Mucha: Absolutely. That’s right.
Ogo Ezeofor: Yeah.
Kirill Eremenko: Awesome.
Ogo Ezeofor: That was very important.
Kirill Eremenko: Thank you. Ogo tip number three, your final tip for today.
Ogo Ezeofor: Let’s see. So I would say make it look nice. Make it look pretty.
Kirill Eremenko: Make it look nice.
Ogo Ezeofor: So, I kind of compare it to of culinary arts or like being a chef, right? So like you can cook the most amazing meal, but if it’s not presentable and if it doesn’t look nice then people, it’ll be kind of hard for people to want to consume it.
Ogo Ezeofor: So I always try to spend a lot of time making it look nice or making the charts look very clean and very well put together and just the overall dashboard, the look and feel of it, the user interface and the user experience as they’re drilling down, what colors am I using, being subtle with things that are not important and then using more vivid colors for things that stand out or areas of opportunity and not overdoing color. Right? So just being able to create a very well put together dashboard that looks nice is also essential.
Kirill Eremenko: Fantastic. Thank you. Love that tip. Absolutely agree. I was cooking dinner yesterday and then I was like, “Oh this rice looks good, but something’s missing.” So I took a tomato and I cut it up like a crown, that’s a way you kind of put tomato for the spiking way put on top of that, oh much better. The rice didn’t taste as good but it looked great. All right Jonathan, your final tip for today.
Jonathan Mucha: So final tip because you guys have covered how to make beautiful dashboards very nicely. I’ll say, think of the psychology of your clients. So is what you’re doing going to somehow make your client feel inadequate or scare them in some way? Are you presenting information that’s going to contradict their beliefs and what things are happening. Those sorts of situations end up with you doing an awful lot of work and with a really unhappy clients.
Jonathan Mucha: So if you don’t take the time ahead of time to sit down with them and say, “This is what I think is going on. What do you think is going on? Do you see a pain point here and is it painful for you or is it painful for the company?”
Jonathan Mucha: If you start … you have to really think about your client’s mental state before you start showing them anything because it’s really easy to embarrass them in a way that you don’t know that you’re doing. Does that make sense?
Kirill Eremenko: Absolutely.
Jonathan Mucha: That’s I think just a key to doing any type of consulting, internal, or external, whatever. You really need to spend the time to understand your client’s state of mind before you get even into the process that Ogo outlined.
Kirill Eremenko: That’s really cool. Thank you. Ogo, have you been in situations like that where you potentially could have embarrassed the client or like you had to, like basically you can relate to something that Jonathan is saying and that thinking about those things helped you structure the visualization into different way?
Ogo Ezeofor: Yeah, definitely. Sometimes that happens with that data integrity issues. So let’s say the data is… You build out a dashboard and the data is being inaccurately reported and it will make the, company look like they are underperforming dramatically. And being able to catch that in advance and being able to do some data validation prior to kind of submitting your dashboard or showing it to the stakeholder is important so that you can kind of build that trust and you’re not kind of upsetting the stakeholder or just losing that trust. It’s something that you just don’t want to do.
Ogo Ezeofor: And on top of that, I would say that Jonathan made a very good point. Understanding the psychology of the client because if they … let’s say if you create a dashboard that shows, the client is drastically underperforming in certain areas, before presenting that to a team of people, being able to kind of have that direct line of communication with let’s say one of the key stakeholders to just make sure that, this is accurate and this is an approach that they would like to take with the dashboard is important.
Kirill Eremenko: Absolutely. So it’s not about bending the truth, it’s about what is it that the purpose of this presentation is? What is the purpose of this visualization? You might find certain things that weren’t the purpose and that are important to the client to correct, but for this specific audience, they don’t want to, stir the waters. They don’t want to present these insights, they want to work on them first and then improve and then fix the situation.
Kirill Eremenko: This may be not relevant to the specific presentation. I think that ties in very well with Jonathan’s comment about constant feedback. If you have that constant feedback loop with the client, then you’re less likely… You are more exposed to their psychology and you’re less likely to offend them or to go with your visualization in the wrong direction.
Jonathan Mucha: Yeah, absolutely. If you’ve got that feedback you can … and if you stumble onto something like Ogo said that “this is a really sensitive little bit of information I’ve discovered here. Your employees, if the sales department is misrepresenting their numbers, how do you want to handle that? Do you really want to get up in your annual sales meeting and embarrass that entire team or would you like to handle this on your own?”
Kirill Eremenko: Exactly.
Ogo Ezeofor: One thing with that I feel like sometimes some companies just have kind of put their ego aside in a sense and essentially kind of own the fact that they are underperforming in certain areas. And I think with this step two of the phase that step two that we discussed earlier in terms of designing the Mockup visual first.
Ogo Ezeofor: If we’re able to, complete that process, then the stake holder will know that these are the charts and the KPIs that we will be tracking. So being able to catch that in that stage will help prevent, that kind of situation from occurring where you’re displaying something that they don’t want to see because you’ve already addressed that in the second stage.
Kirill Eremenko: Yeah Ogo, that’s, a very good point. Yeah. Follow Ogo’s methodology. Avoid mistakes. Awesome. Okay. Thank you. And my tip number three is get certified. So I’m not saying this just because I’m creating courses on certification, that is the case, but I also found that by taking the certification myself, the first one on Tableau Desktop Specialist, I found that A, it’s very easy. B, it’s not super expensive.
Kirill Eremenko: If you compare it to the Tableau license is like 800 bucks per year, the certifications is $100. It is, pricey, but it’s worth it. And why is it worth it? Because once you’re certified, it’s, you already know all these things, right? You already learn Tableau, you’re already using it, add the certification on top. I don’t know, maybe other tools probably also have certifications, but I’m talking specifically about Tableau.
Kirill Eremenko: Add the certification on top and then you can … you have, a certain weight to the conversations you’re going to be having, whether it’s with your clients, if you’re a freelancer, whether it’s with your employer, if you’re employed, whether it’s, with your interviewer, if you’re looking for jobs.
Kirill Eremenko: Once you have these certifications, it’s kind of a very strong statement that “Hey, not only I believe myself that I know these things and here’s my portfolio, but also I’ve done, I’ve gone the extra mile, I’ve done the certification on my own. I’ve invested my time and effort and money into this.” And that shows that I’m serious about this tool plus, because I have the certification, I know all the things that I need to know at this level.
Kirill Eremenko: The exam is not hard at all. It’s a multiple choice exam, just prepare for it for a few days, or weeks depending on the level of the exam. But I really highly recommend like doing it myself. I highly recommend for everybody if you’re serious about Tableau, get on top of it. It’s going to be totally worth it for your career.
Jonathan Mucha: That’s a good tip.
Kirill Eremenko: That’s including you guys. Ogo what do you have to say?
Ogo Ezeofor: Yeah definitely.
Kirill Eremenko: Awesome.
Ogo Ezeofor: Of course I’ll be looking into that.
Kirill Eremenko: Awesome. Well this brings us slowly to the end of the podcast. I’ve, had a lot of fun. Lets, recap on these tips. So Ogo your three tips, keep it simple and use color, using KPIs effectively. And number three, make it look nice.
Kirill Eremenko: Jonathan. Number one was download kite for Python for code completion. Number two, constant feedback with your clients. Number three, think of psychology of your client or think of the psychology of your client.
Kirill Eremenko: And my tips were number one aggregation and granularity. Number two you present, your dashboard supports you, if that’s a presentation in person and number three get certified. Yeah. Had a great time chatting with you guys. How did you guys feel about this podcast?
Jonathan Mucha: There was a lot of fun. Kirill, thank you much for having us. It’s been so much fun.
Kirill Eremenko: That’s my pleasure. Absolutely. My pleasure. Some final notes before we go. Where can our listeners find you connect with you, get in touch? I don’t know, like with Jonathan maybe somebody might want to get to know a bit more how you use visualization with machine learning and integrate those things sounds really cool. It sounds like you have some great ideas to share.
Kirill Eremenko: Ogo, you have your freelance company. Maybe somebody might want to engage you for your services. What are some of the best places to get in touch?
Jonathan Mucha: So LinkedIn for me is always a good place to get in touch with me. So you can reach out to me. LinkedIn, Jonathan Mucha. M-U-C-H-A. I should be one of the only guys.
Kirill Eremenko: Got you. And Ogo?
Ogo Ezeofor: Same for me. You can reach out via LinkedIn Ogo Ezeofor on LinkedIn and also ogo.ezeofor@outlook.com and then I also have a website ogoezeofor.com as well. That should be launching shortly.
Kirill Eremenko: Fantastic. And Ogo, you mentioned you are looking for an expert data visualization artists, somebody who knows visualization somebody artistic to, kind of collaborated with in your freelance career. Tell us a bit about that. Maybe there’s somebody on the podcast listening who might be interested in that.
Ogo Ezeofor: Yeah, definitely. So I’m always looking for individuals who have great ideas in terms of how to build dashboards, a more modern looking dashboards that are very … kind of unique and using some artistic abilities to kind of create a very creative dashboard. So kind of that hybrid of a kind of like an artist and an artist who has like technical abilities, that’s kind of what I’m looking for in terms of data visualization resource.
Kirill Eremenko: So Jonathan, you said … you mentioned before the podcast that you are looking to transition into the space of data science and visualization. Congratulations. That’s, a huge decision. Very, excited for you and tell us a bit about that. What kind of a role are you looking for? What kind of a … What excites you the most in data science or visualization these days?
Jonathan Mucha: Yeah, thanks Kirill. The thing that I really discovered when I was actually at Matt Dancho’s presentation was providing business impact is so much fun. I’ve been doing data science kind of forever and I just didn’t know it until about 2017. So I’ve spent a lot of time, we built a startup, I’ve learned a lot about business. I’ve learned a lot about how to solve problems and looking to put all those things together and provide some impact for some really progressive companies looking to have a data scientist on their team.
Jonathan Mucha: So I’m going to be doing a mentorship through SharpestMinds with Ayodele Odubela, which is going to be really fun and I think help really build up my portfolio. And so after that I’m really going to be out there looking for data science opportunities in probably the business world. I really want to focus. Yeah, that’s it. I really want to focus on being able to provide meaningful impact.
Kirill Eremenko: That’s awesome man. Congratulations and best of luck on that journey.
Jonathan Mucha: Thank you.
Kirill Eremenko: My pleasure. I also want to say that it’s really cool. I love these moments. Are very rare but I know Jonathan, I’ve seen you, we met at DataScienceGO, I’ve seen how you can help people. How passionate are you about this. I know how active you are in the SuperDataScience community, you’re sharing your expertise there and ideas.
Kirill Eremenko: But what I mean like these moments are so rare is that, there’s so many companies out there looking for talented, passionate, driven data scientists. And at the same time there’s so many people who want to be data scientists, who want to be passionate, who want to be driven, who, want to have a career.
Kirill Eremenko: But the thing is that not all of them are putting in the effort, the time. Not all of them, they want to be passionate, they want to be successful, they want to be driven, but they’re not … they rather, do something else or go to bed early or not wake up. You know what I mean? It takes effort to go through these courses to learn these things to go fly across the country to attend a conference in California, if you live in New York. So I know about you that you’re really like, you’re one of those guys that if companies knew you’re there, they’d hire you in a heartbeat. 
Kirill Eremenko: And why I love these moments is because like you’re on the podcast and there’s like 10,000 people listening to this. So inevitably there’s companies in New York, Upstate New York area, so my call to all the companies in whoever’s listening from Upstate New York or New York, if you need a talented passionate data scientist, Jonathan is your guy and he’s probably not going to be on the market for long enough after this episode. So hit him up as fast as you can.
Jonathan Mucha: Well, thank you for the endorsement, Kirill that means a lot. You’re absolutely right. You know what I sort of joke with people, I tell myself that I really want to learn French and I’ve really wanted to learn French for probably 20 years. So yeah, I get it. Right. There’s a lot of people out there that are, in the same spot. They’re like, I really want to be a data scientist, but I also have all these other things going on and, it does, it takes dedication. I get up at four o’clock in the morning most days so that I can work on projects before I go to work.
Kirill Eremenko: That’s great. Very inspiring Jonathan. And thank you for sharing this and I’m sure there’s people out there listening who are going to be, banging on your door in no time.
Jonathan Mucha: That’s exciting.
Kirill Eremenko: All right guys. Well, on that note I think, we’ve gone way over time is probably one of our longest podcasts out there, so I just wanted to say a huge thank you for coming on the show. I had a great time chatting with both and the insights that were shared the knowledge bombs that were dropped. I think people are going to be getting so much value out of this, so thank you so much.
Jonathan Mucha: Thank you for having us Kirill. It’s been so much fun.
Ogo Ezeofor: Thank you Kirill. Really appreciate it.
Kirill Eremenko: All right guys, take care.
Ogo Ezeofor: All right.
Kirill Eremenko: So there we have it. So that was Jonathan Mucha and Ogo Ezeofor. I hope you enjoyed this episode as much as I did and got as many cool takeaways from this podcast. I really enjoyed sharing the three tips with the guys. That was really cool and it was interesting to see how the tips we shared were slightly different, Ogo focused on keeping things simple and making visualization very digestible.
Kirill Eremenko: Jonathan focused more on a psychology and talking to your clients and I talked about like, I don’t know, different fundamentals of visualization from a aggregation and granularity and how to present and also the certification aspect of it.
Kirill Eremenko: Probably my favorite takeaway was Ogo’s methodology. That was a really cool, something that’s like I am not used to doing, approaching visualization, not looking at the data at the start. It’s very tempting to look at the data that you have to and then include structuralization based on that and his approach is backwards or his approach is upside down compared to that. 
Kirill Eremenko: So basically start with the questions, then come up with a design, interact and then original dashboard. I really enjoy that. So huge shout out and thank you to Ogo for sharing that. And of course there are plenty of other takeaways here and if you want to get in contact with the guests as usual, you can find their profiles or URLs to their LinkedIn and any other contact information in the show notes which are available at www.superdatascience.com/319, that’s www.superdatascience.com/319. Make sure to connect with Jonathan and Ogo on LinkedIn.
Kirill Eremenko: If you are looking to work with visualization company and visualization freelancer hit Ogo up. Or if you’re looking to collaborate on visualization project with an interesting freelancer who’s already got processes set up, also hit Ogo up.
Kirill Eremenko: On the other hand, if you’re looking to hire a visualization expert or a passionate and talented data scientist who I personally know and you’re based, your company is based in New York of Upstate New York or it’s an online remote company, then make sure to hit Jonathan up.
Kirill Eremenko: And as usual you can also get any materials we mentioned plus the podcast transcript in the show notes as well. And some final thoughts. Look out for the Tableau certification courses, which I am recording. Well we’ll be recording in the next month or so and, which will be released in January and February next year. So look out for those. I would love for you to take them to be part of them. If you took the Tableau A-Z and the Tableau Advanced course, you know that they’re fun and that there are some … we’re going to be some really cool case studies, hands on experiences, exercises that you can do and plus then you can definitely with much more certainty pass your certifications.
Kirill Eremenko: The ones that I am doing myself right now, so I’d be very excited for you to join in those courses. Look out for those, look in your inbox and look out for any announcements we send around those in January and February.
Kirill Eremenko: And one final thing I would like to ask of you today is if you know somebody who’s passionate about data visualization, who is interested in the field, who loves Tableau and other visualization tools, then send them this episode, share the love, share this podcast. Let them also learn from these insights and get some valuable takeaways from here. They’re easy to share. Just send them the link www.superdatascience.com/319.
Kirill Eremenko: On that note, thank you so much for being here today, ladies and gentlemen, can’t wait to see you back here next time. Until then, happy analyzing.
Show All

Share on

Related Podcasts