See and understand your data

Geraldine Zanolli: Thank you. Thank you again for having me, and welcome And to this talk, see and understand your data. I'm Geraldine, but everyone is calling me Gigi. I'm a developer advocate at Tableau. And just to give you a little bit background about myself because I think it was going to be useful for this presentation, Actually, I started at Tableau as a presales consultant.

So I was working with EMEA. I was based in London. EMEA customers to adapt and use Tableau in the right way. So when I joined the dev well the developer relation team at Tableau, I really tried to apply all of the learning and all the advices that I was giving to customers to DevWell and developer relations. So it's why I decided to put this talk together.

So we are going to talk about why data visualization, why data visualization is important, and also why is it important for DevWell. Then we will start to talk about where to start once you have a blank canvas. What is the best way to start? Which data can you use and how to visualize it? It's why I'm giving you tips and tricks for how to visualize your data and any advices to continue and apply these skills in your team organization.

So at Tableau, we have one mission. We help people see and understand the data. When we say people, it's anyone that is a business users. That then you don't need technical skills to get started with Tableau. So I don't know how many of you are familiar with Tableau already.

And if you are familiar with it, you might be familiar with Tableau Public. It's our kind of YouTube of visualizations, and we help people build awesome visualization that they can share online publicly. And you might be wondering, oh, they are doing a lot of visualization. Do they have API? Why do we have a developer relation?

So we have a bunch of APIs just to give more context about our developer platform tools and APIs. We go from automating task on the server to connect to data sources. But that's not today's focus. Today's focus is why data visualization. It's important to start with this game, I think.

So that's in a way you should look at your screen. And now I will ask you to count the number of nines. You can press it on Slack and say how many nines nine can you see on this slide? Give you few seconds there. Now you might have a number in mind.

How confident are you with this number? Are you like 100% sure that's my number, that's a number of nine? And now if I go to the next slide and I ask you again to count the number of nines. How many nine? Can you see?

Now you have an answer. You asked way quicker than before because we added the color there. And it's another part of your brain that is actually using and say, oh, okay, color red, nine, I can count it. And also, how confident are you with this answer? Way more confident than before.

It's why data visualization is important. Let me give you another example. So at Tableau, we have this superstar. It's called fake, a supermarket, but it's selling store, it's selling a bunch of different furnitures and accessories. And you look at this base and you're like, okay, that's bad.

Like first thing in your brain is like, something is going wrong at Superstar. And you say, okay, what is it about? Is it because of a graph? What if I turn the axis and change the axis? Straight, you're like, okay.

Are they losing money? And I'm sure why now? It's just not sure. What about if I change the color? Adding this blue color there, it's way more positive message.

You have a better insight and you're like, okay. There's are they losing money? Like, I'm not sure anymore the color say something else. And now we change the title and surprise or profit is going up. So you have a way more positive message based on this visualization, and you have a different message send across to your audience.

If we put these two slides, two visualization side to side, it's the same data that represent the different ways and with different context there. But why is it important in DevWell? Why do I care about how data is represents? So first is having a better understanding of your developer audience. You have quick insight in what is your developer audience, why is your developer audience cares.

You can create them more targeted content for your own audience. And of course, I think it's an issue, in DevRel, is that you have to prove your value, your work to other teams or executives. You have to say, oh, that's my work and that's the impact I had. So that's a great way to show your value to other teams and other parts of your organizations. It also help you to define your different engagement points, which one is the most relevant for your audience.

On that, I put a framework together when we talk about data visualization. I like to start. Of course, I put the avocado back on the slide. So first, ask the questions yourself. You say, what do I want?

Which data do I want to see? What are the insights? Then once you have this question and goals, question can be goals as well, say, okay, I don't have this specific goal and I want to answer these specific questions. Where do I find the data? How do I access the data?

And only once you answer these questions, you get the data and have the question line up, you can visualize it. And you will see if you are new with data visualization, but once you answer one question, you have like thousand of of new questions coming up. You're like, what about that? And what about this? And then you start again.

It's just a circle. And you ask new questions, find new data. So let's define start by asking new questions. Ask finding new questions, define new goals. So what are your goals?

I think you had multiple talks already on Tuesday that I well, watch again as we can really recommend define your goals and your metrics. But the summing is where are you in your developer journey at the company? So Tableau developer program has been around for two years. So we went through the beginning of the developer program where we created our awareness. It's like, we are here.

We are there. We're helping you, helping developers to the adoption of APIs and the program. Then you go for the growth. You want to increase the usage of your APIs. Then you should align with the goals of your department.

You have different goals if you are in marketing, in dev, or in sales, for example. And defining your own metrics is the start of your journey with data visualizations. You have a blank canvas and you have questions to answer. Something I noticed as well is that you have I have, at least, a tendency of having a thousand of your metrics. And I have it's like, want to look at that and that and that, especially being a data geek.

I'm like looking at thousand of different metrics. And I think it's important to say that if you have too many metrics or no metrics at all, you have a tendency of look every looking everywhere. And you don't have a real focus on what are your goals and what is what you want to show to your executive team or to any else in the organization. So now you have your matrix and you have the questions you want to answer. The next step is finding the data.

So the most word, I mean, there's another way to say data is a new bacon. I don't know if you heard that before. But, yes, data is the most valuable resources right now. And data is everywhere in DevRel. You have social media blog post that you are posting or external blog post.

Are you abusing bugs? How many forms? You have thousand of different data, and they are out there for you to collect them. So find new goals, define also the data source you can use from bug recording to get a pull request or bugs waste. The next step is going to be understanding where your data are.

A bunch of social media, like Twitter or I don't share about LinkedIn, actually. They have APIs where you can get data. Blog post, you might have a tracker code of to have Google Analytics data. The project logs and the forms might be in your back end data sources or store in Excel spreadsheets. Find where are the data and how can you get access to them.

And once you have all this data source set up and you align, you have okay. I have 10 different data sources. I encourage you to take a step back and don't think about one singular data sources. What about combining different data sources altogether? If I tweet more, do I have more views on my blog post?

What about the bigger picture? By combining these data sources that are separate, they are like APIs, that's back end. You can combine them to have the bigger pictures. And trust me, it's not the fun part. I went through that and I felt like this guy because I was struggling.

It's like the data source was not I'm just going to pass the next one so you don't think. It's too much animation on one side. But it's not the fun part. You're going to get the data source and you have to clean the data sources, get the admin right if you don't have. Or do some digging and say, oh, I have this bug pause, but I don't know if I have a Google Analytics track.

So that's the part where you actually have to challenge yourself or your team to get access to the data and clean the data so you can actually visualize it in the right way. So now you have your goals. You have your questions, your metrics. The next step is how do I visualize it? And I'm going to give you some tips and tricks that are valuable to visualize your data and get your message across.

It is important to have clarity in your dashboards. Right? If you are building a dashboard for yourself or for your team or, let's say, for any of the executives, Do they do you want to tell them a story? Say, hey, if we have a blog post, we have more views on the GitHub accounts. And that's your story.

And you don't want the person looking at your visualization to be, okay. So I have this dashboard. I need to do some insight. I'm going to find my own insights. That's the data and they are visualized.

So if you want to let users having finding their own insight on the data, be clear that you are telling it in your visualization, in your dashboards. So use of the drop down menu or filtering, like, depending of what you want how you you user is how you want your user to use your this. Color is important, and you might start with a unicorn puke because you are getting real into it. You're like, I see. It's like, I can see that.

I can see that. And, like, you have all these colors because you have so many information that you want to share, and you end up with a lot of colors on your dashboards. And it looks like a uni a rainbow. And you're not if you look at the right dashboards, no message across. You have a lot of data and the person looking at your dashboard will be like, okay.

That's too much for my eyes. I can need to stop. So use the color wisely as a weapon. Again, which message do you want to get across? We saw it with example at the beginning, the red color is negate is associated to a negative fault.

When the blue color is more neutral. I like to start with example, my dashboards with a gray color, so an add color at the end, so I can really get my message across. Text is important. So it can be filtering the labels. The title, we saw it again at the beginning.

The title is what which message do you want to come across. And I think I'm I mean, I've used it if you notice. I'm not an English native speaker. So sometimes I send dashboards to my manager, and he's like, oh, I'm not sure what you're trying to say in this dashboard because I'm using Funglish as a new language that I made up. And it's not the same message you're sending across even if you have a data when someone doesn't understand your caption or your title.

I think another mistake was when we start with data visualization is trying to make the most of the visualization and get attracted to visualization that are not that impactful, but they are pretty. Like this bubble chart is really pretty. Which message is it sending across? We are not sure here. It's a lot of colors, it's pretty, it look like a flower, but you're not sending a message across with pretty visualization.

You need to think about the visualization you want to use before and be like, okay. The bar chart is not the prettiest one, but it's efficient and get your message across. Another advice, especially in DevRel when we are back to normal, we are always on the road on, like, between two flights, and we don't really have time to open the laptop to look at the lat metrics. So think about mobile and how someone with a mobile is going to see your visualization and your data. You don't want them to scroll and or to, like, zoom in on the mobile.

It's not a great experience. And they might just close with dashboard. I don't want to use it. Now that I shared advices on how to start with data visualization, I think it's also important to show you how we use it at Tableau. So on the left side is data where we track our dashboard we use to track our hack at our events.

It's a really simple dashboard that it does the job. We are looking at the number of registration of the time so we can check it when we need it and when we give it to executives. They, okay, we have that many people coming. That's how the different companies they are coming from and from which region in the world. It can be, okay, we had a hackathon.

We were supposed to have a hackathon in Tokyo at the beginning early this year. But you can see that it was a gold ball of hackathon. People are actually flying over, so we had to work with Visas. And once once we had to cancel the hack, it was also useful to come back to Visas as well and say, okay, we have to send emails to that of that people, that many people. The middle one is the most used dashboard in the team, is the developer program stats.

So we are looking at how many developers we have in our program. But not only that, we are also looking at the developer sandbox sites. Once you join the developer program, you get a sandbox site. So we look at how many people activated the site. And that give us an insight of how many people are actually active.

And recently, single view of the peak is when we launched the mini challenges, because we saw that many people were activating the sites. So we launched mini challenges to encourage them to use our sites. We also use data visualization to communicate to developers. The right form is a data visualization on Tableau Public, the YouTube visualization. And we noticed a gap between the business user and the person coming to the hack.

The business user, they have a lot of ideas of what they want to see at as a hack project. And the developer might come to the hack and say, okay. I have all these skills, but I don't know what to build. So they can submit ideas on the view on our this dashboard on Tableau Public, and people can avoid for ideas. So hacker can go in and vote and see before starting the hack.

Okay. What are the business user asking? More example is, like so this one on the left is also tracking our events. We have a monthly event that's a screen demo. And we are looking at, again, people over time.

Do we have are we increasing the number of people attending? And I think the most important that we added on the top on the bottom left is how many people are coming back. We are doing all these events every year, every month. Are there people coming to one event and then they don't come to any other events? And on the right side is a survey data where we combine survey data and submissions from our data that have many challenges.

And just by looking at these dashboards, one insight we took is that the first challenge the first challenges didn't have any real instruction. It was like, try to build that. And even if it was easy enough, we needed to add instructions. And we added in the second challenge, and we saw a growth of term of users participating to the challenges. So it's how we improve the challenges as well.

So how can you get started now that I showed you all the best practice and share this kind of framework we have? I think it's important to have dedicated time for your presentation for visualization for data visualization. And that's an example. My manager has this new initiative to get us more hands on with the data. So he created this slide both saying, hey.

Remember, you you need to spend one hour exploring the data and share your insight as a team as well at the end. So every week, we had one hour dedicated to just exploring data and find new finding new insight, and we were presenting to the team. It's why it's important to work as a team because you might have different insight and different goals and different background as well. And sharing this perspection perspectives and sharing new insights with you might have one insight and someone looking at the dashboard would like, oh, I actually see another insight. And and that's just interesting to bring all the people on the call right now or in the room in the future.

So I think that's it. And I hope you enjoy this presentation and you get inspired to do more with your own data. And you have, yeah, you're going to come up with great visualizations and to analyze your level data.