Right before the holidays, I had the opportunity to sit down and have a nice, long conversation with Bill Lyons. He's one fascinating guy - hugely into data (not surprisingly), extremely curious, and an avid bowler! You may know him from his Day-night map document or from him helping people throughout the community. Take a moment to read through this interview, and learn more about Bill and his Tableau story.
(Oh, and he's a big Star Trek fan )
Tracy: To get started, my first question is how long have you been using Tableau?
Bill: Since June of 2013, version 7.
Tracy: And how did you discover it? Was it through work?
Bill: Well, interestingly, back prior to 2013, I was very much a Microsoft person. A couple of us were trained in ProClarity, right around the time that Microsoft acquired them. After the acquisition, it wasn’t anywhere near as user friendly or powerful as it once had been. They integrated it all together into what they called Performance Point Server, which they then rolled into SharePoint, and it didn’t resemble ProClarity at all. It took a lot more time and effort to build anything. I was getting more and more frustrated with Microsoft. Every time a new version, you’d have to upgrade all of your servers to that version, and it’d take years to get through red tape, and so forth. Then when you did, it was a completely different paradigm because you’d have to learn all new software. When they decided to replace everything I knew with PowerView, I had had it. It was the last straw.
I went to a Gartner conference and was talking to Gartner consultants, trying to figure out what direction we should go from that point. I had heard about Tableau before that, one of my colleagues had tried to push me to use Tableau, but I was basically married to Microsoft. I wasn’t ready to do it. At this point, though, I was considering everything. At the Gartner conference, I happened to walk by a session that I hadn’t even planned on going to, but they had a free lunch. So I thought, hey, a free lunch! Why not? I’ll just go in, eat, and listen. It was Andy Kriebel, who was at Facebook at that time, working with Cloudera and Facebook, talking about how they were visualizing petabytes of data using Tableau and Cloudera. I was completely blown away. I saw all the things I was looking for – the ability to do mapping of locations of stations and donors, and the ease of building the views. I determined that when I got back to the office, I was going to start the evaluation of Tableau.
Once I got back, I dedicated a couple of weeks of time to Tableau. I thought if I’m going to do the two week trial, I’m going to get the most out of it. I allocated my entire two weeks to figuring out if Tableau was the right product. I was able to rebuild all of the things that had taken me years to build in Performance Point, plus some things I couldn’t do before, I built in those two weeks in Tableau. I was so impressed, I went to my boss and said, this is what I’m doing, and he said “Fine!” and that’s the story of how I got into Tableau.
Tracy: That’s awesome! Were you working at Educational Media Foundation at that time?
Bill: Yeah, I’ve been here 16 years.
Tracy: Wow, you’ve been there for quite some time. What specifically is your role?
Bill: My title is Principal Data Scientist. So, we do data science. The name of my department is Data Science and Insight. The insight, I believe, is really key. I have a couple of slogans up on my wall. One is “Without data you are just another person with an opinion.” And on the other wall it says, “Data without insight is noise.” I very, very strongly believe that. We can produce dashboards, reports, and all sorts of stuff, but without understanding the meaning of it – it’s just noise, and it confuses people. That’s really a big part of our job, to look at the data, and look at how we can make positive changes in management decisions and so forth so that we can all accomplish something.
Tracy: It sounds like you use Tableau on an everyday basis, is that true?
Bill: I personally, am not in Tableau every day, like I used to be the first couple of years. Now, I’m in it probably 2-3 days a week. Since about March, I’ve been spending a lot of time in Alteryx because a lot of our challenges, now, are with pulling together data from different data sources like Redshift, Excel, and other servers and systems that had never been connected before. So Alteryx is helping us do a lot of things that we hadn’t been able to do. We do some direct connections in Tableau from Redshift and with other servers, but the data needs to be transformed in some way. Matching account numbers or some other way that we don’t have a direct, complete join like Tableau would need to do. Tableau is great, especially with cross-database joins, but that assumes you actually have a field to join on. So we need to do some manipulation there, and Alteryx has helped with that tremendously.
Tracy: Had you ever tried to use these data sources together before?
Bill: Oh no, we never even attempted it before. In fact, we were trying to do some things with Tableau (of course, this was pre-version 10), so we were just using blending. We got some things done, but Alteryx is very, very strong with the geo-spatial functionality so we do a lot of work with those for our transmitter coverage patterns, and things like that. We export those to extracts that we can consume with Tableau. So the two work together very, very well, and we do that a lot.
Tracy: Do you use Tableau in your personal life or in personal projects at all?
Bill: Yeah, interestingly. One of the projects I’ve been working on since September is a combination of Alteryx and Tableau. I’m using Alteryx to do a screen scrape for the website of our bowling league to take all the bowling scores for all the players in our league and compile them all together, load them into an Excel file, and bring them into Tableau. I’ve been visualizing what pairs of lanes people perform better on, what pairs of lanes people perform poorly on, is the winning team a higher average or lower average than the losing team in each matchup, and comparing different handicap methods, various things like that. It’s fun.
Tracy: That’s so fun. Have you found anything interesting? Do you know which lane is the lucky lane?
Bill: Well, lanes 15 and 16 appear to be the worst, and 13 and 14 appear to be the best. More interestingly, one of the things I’m finding relates to handicapping. The USBC (United States Bowling Congress) has a document on handicapping methods that says the higher average team tends to beat the lower average team most of the time even with a 100% handicap. My stats are saying that is completely not true. Their methodology is based on whether the average is higher or lower than the median of the league. I think that’s a fallacy because your matchups are not playing against the median of the league, your matchups are playing against another team. And it’s a matter of how your two averages compare and not how they compare to the median. I’m going to wait until the end of the season to see if my numbers hold up, and if they do, I’m going to write a paper. I expect Tableau to help visualize that.
Tracy: So cool! I never knew you were so into bowling. How long have you been into it?
Bill: I started back in ’95, then I quit for a while. When we moved, we were near a different bowling alley, so I joined that. I’ve been bowling continuously since maybe 2008. I’m not that great, but I’m the league president, so when the league cooks up some stupid rules, I’m using data to try to convince the league to change them next year.
Tracy: I love it! What type of data interests you?
Bill: Anything I can find that can make a difference. If it can be used to make a decision or prove somebody wrong or increase productivity, if it can make a difference, then that data matters to me. Ironically though one of my favorite projects or vizzes - the day/night mapping viz - was more informational. That came to me at time when I was frustrated with work, and I didn’t feel like going to work in the morning, and that project re-energized me and got me going again.
Tracy: Sometimes you need something like that. I like that response – it’s important to be able to do something with the data, not just look at it for the heck of it.
Bill: Yeah. We’re a donation based organization, and one of the things I found was that the long-term value of a $20 a month donor was far higher, I mean 2-3 times the amount, than a one-time $500 donor. So if a person gave $500 one time, over the next ten years, they would actually be giving less than someone who was giving $20 a month continuously. We included attrition rates, inflation rates, credit card decline rates, likelihood of giving another gift, etc. to calculate the long term value. Once I was finally able to present that visually to our executives, they changed the way they handled the pledge drives on the air. It has made such a difference in our income that I had our finance guys coming to me, asking “What’s wrong with my budget? We keep going over budget on our income by 10%. What happened?” And I said, “Well, that’s what happens when you listen to the data.” It’s made a several million dollar change in our income.
Tracy: That’s really impressive. When you stop to think about it, it makes sense. $20 a month doesn’t seem like quite the burden that $500 does (even only if it’s a onetime payment). And if you do the $500 payment, you feel like you’re covered for a while. Ok, next question, what’s your favorite Tableau project you’ve worked on?
Bill: The day-night map (https://community.tableau.com/docs/DOC-6573), which was spurred by a forum question where the person just asked about simulating a day/night map. Their question referred to a website that had a map that would show light area for day, and a darker area for night, and it had a hard line for what is known as the terminator (where the sun rises/sets). That piqued by interest because there was a lot of trigonometry and math involved, and I thought, I can accomplish that. At that point, I was trying to make my mark in the Tableau community, and I thought – I can kill this one! I got so excited about it – I worked all night long that first night. By 9 in the morning, I had something I published with that first version.
But it still bugged me because it had that hard line – and we don’t experience that. It’s not suddenly dark. It gradually changes. It bugged me for a while. How do I portray that twilight transition? I did some research, and I posted that picture of the white board that had all that trigonometry on it, and I puzzled over it for quite some time. Maybe 2-3 weeks. The hard terminator line made the trig much easier, because the point where the sun is crossing the horizon, the elevation angle is 0. In trig, when the angle is 0, a lot of terms of the formula cancel out, making the math much simpler.
However, for twilight, the suns elevation angle is not zero. According to Wikipedia, there are three stages of twilight: civil, nautical, and astronomical. And each of those are 6 degrees below the horizon of the sun. When you’re 6 degrees below the horizon, or a solar elevation angle of -6 degrees, those terms don’t cancel out, making it much more challenging to calculate the exact latitude and longitude for any given time of day. So after staring at it on my whiteboard for a long time, I realized that in order to solve that I needed to change it to a quadratic equation. Then, each term became a trigonometric expression, so I was able to solve it mathematically. But that still didn’t solve some problems. When you have square roots you have to take both the positive and negative number, and so you end up with some values that are either impossible, such as a latitude that is greater than 90 degrees, or you end up with two potential locations on the earth, and they can’t both be sunrises. It’s impossible! And you don’t know which one is valid.
So I had it working for certain times of the year, but when it came close to the equinoxes or solstices, things became a little weird. If you look at the day/night on a Mercator projection, which is what Tableau does, it looks like a sine wave. Except when you get to the equinox, where the sun is right over the equator, that sine wave can turn into a circle because both of the poles are getting sunlight. So that got rather challenging. And at the transition from being a sine wave to being that oval shape, how does that transition happen? The curves had these weird tails that didn’t make any sense. A friend of mine who has a PhD in satellite earth science was helping me figure this out, and he said to me “I wish we had a way to animate this.” And suddenly a lightbulb came on in my head – Tableau can animate that!
So I threw the solar elevation angle onto the page shelf and it animated it, and we could see what was happening in those transitions. Suddenly, we realized that all we needed to do was invert the sign on one of the terms in the formula because of the way quadratic equations work, and everything lined up perfectly! I illustrated that with Story Points and posted on Tableau Public. That was the only time that I’ve found where the animation actually helped us solve the problem.
Tracy: You spent so much time on this!
Bill: I spent about two full weeks on this.
Tracy: So is your background in math?
Bill: Yes and no. In high school, I was the top of all my math classes. I took some college level classes in high school, but once I got into college, I got into computer science and broadcasting which is the industry that I work in, and I had a really lousy calculus teacher. So I lost interest after that. But this project resurrected a lot of the old high school memories of trig and analytic geometry. It was very stimulating to me.
Tracy: What is your favorite feature in Tableau?
Bill: There are so many. I love the simplicity of drag and drop. That was one of the first things that attracted me. That and the ability to geographically map data. Those were the initial things.
Every new version that comes out there’s something new and better that is my new favorite. You know, what was my favorite in the beginning, you kind of take for granted after a while, and there’s another new favorite. LODs were huge. I still use them a lot. I think my new favorite that’s coming with 10.2 is the geo-spatial features. Just depends on what year and what version it is.
Tracy: What’s your biggest feature request?
Bill: Hmm, there were so many that they talked about on the roadmap at TC. I think the most useful for the upcoming future will be the things that are more collaboration and social functionality. The ability for people to make a comment on a mark on a Tableau Server view so they can ask a question, get answers, and so forth. Tableau already makes the data come alive in being able to interact with it. Having one more level of interaction where not only can you interact with the data, you can interact with another person about the data.
Tracy: So it sounds like you have Tableau Server then?
Bill: Yes, we do. I do some of the administration – like organization of the projects, etc., but someone else does more of the technical administration.
Tracy: I see. Ok, time for some community questions. How did you discover the Tableau Community?
Bill: I’m trying to remember. It seems like there was a problem I had and I wanted to know the answer to it. I’m more of an introvert. I like to find answers to questions myself, so I went to the forums just to look for answers. I wasn’t looking to actually offer anything to anybody.
The first thing I posted that was of value to anyone was the mapping polygons and points together on a dual-axis map. That was because I had to do a lot of research to figure out how to do it, find an obscure video that had been posted from an early Tableau Conference (which I still haven’t been able to find, since). But because it was so hard to find, it was just a video, no documentation, I decided to document it. That was the first thing I contributed of value. Even yesterday there were questions and comments on that.
My initial thing was to look for answers to questions, like how do I do this? When it became really hard to find that answer, I decided to publish it so that others could benefit from it too.
Tracy: Yeah, what goes around, comes around. What’s your favorite part about the Tableau Community?
Bill: I have been amazed at how it has grown and how helpful people are. I’m also a little bit active in the Alteryx community, but it is nowhere near like the Tableau Community. It has become competitive, who’s going to answer a question quickest. It’s getting harder and harder to find questions to answer because there’s just so much of that competition. I really am amazed by that. I like the supportive group out there. You can get an answer to most things if you write your question clearly and concisely. People will respond, and I think that’s amazing.
I encourage my staff to do it too. I think it’s really valuable to people to help others. You gain by helping someone else. You learn more by trying to answer someone else’s question than you do answering your own question. When you answer somebody else’s question, not only do you learn the functionality, but you have to explain to them why and how it works, and document it. That cements it in your memory, and I think that’s really important to the learning process.
Tracy: What advice do you have for new members of Tableau and/or the Tableau Community?
Bill: The most important thing is documenting and thinking through the problem as you’re writing your post. One of the most frustrating things is when the question is so confused, you don’t even know how to answer it. So my advice, then, is to maybe try out your question on a colleague or something so that you know your question is clear, and make sure you’ve made your best effort to solve the problem before you post it. Some people give up and post their question so early in the process, they haven’t even started to give it a shot.
As far as new users, my biggest advice is to explore. Watch every one of the videos. Tableau has so many good videos out there. I spent my first few months watching so many of the videos, and not just watching them once, but trying to do what they did, and pausing and rewinding it until I really understood what was happening and was able to accomplish it. And the forums are a great source for resources. Don’t hesitate to go in there and see what others have done. Most questions have already been answered many times over if you just do some looking.
Tracy: My last question is – is there anything else you want the community to know about you or you want to share?
Bill: There’s a book called Weapons of Math Destruction, and it’s about how data is an extremely powerful tool, and it can be used in very good and positive ways to benefit society to benefit your organization, but it’s also a very dangerous tool that can have negative impacts in society or your organization. We need to be careful about what data we share and what conclusions we draw from it. It’s a very powerful message in using data in appropriate and beneficial ways.
Bill's Profile: https://community.tableau.com/people/bill.lyons
Tableau Public: http://public.tableau.com/profile/bill.lyons