For the past two months I’ve been connecting the dots, albeit at a much larger scale, here on the Data team at Bump as an Intern. In essence, what we do on the Data team isn’t all that different from that Connect the Dots book I played with as a child. Sure, we might trade in that book for a multi-terabyte server of log files and our crayon for a text editor but the essential elements remain the same. We use data to create understanding and stories that shape how Bump and its underlying technology evolve.

I’ve always loved building things, understanding how and why things worked. Out of the gate on Day 1 I realized that Bump is very much data driven. We use data daily to build and iterate extremely quickly. If we make a change to our app, within minutes of release on the app store we’re able to tell if a feature is popular, causing confusion, or broken.

 Three Things I Love About Bump

Massive Scale: Obviously when you’ve building something for 45+ million users you have to have some consideration at the back of your mind of how things are going to scale. While this might be one of the most challenging aspects of what we do it is also makes it exciting.

Culture: One of the things that really drew me in was the amount of passion around building an app that people loved to use. Between our two week sprint cycles and Hackathons this quickly became apparent. An interesting side story: although I had already been using Bump for quite a while it was a post by Jake on Hacker News that got me to where I am today.

The Team: One of the smartest and down to earth teams I’ve been a part of. I think one of Dave’s favorite quotes comes from a TED Talk Simon Sinek gave,  “It’s not what you do, It’s why you do it”. I think it sums up our company ethos quite well.

A World of Bump

One of the (many) cool projects I’ve been working on is building interfaces for understanding the distribution of Bump users and how they use the app. We have millions of location data points that travels through our system on a weekly basis which I figured would make an interesting starting point to better understand our users and get my feet wet with our data.

There are dozens of ways to map geolocation data to a map with R, Fusion Tables, and Tableau just to name a few. I tested a few before settling on the Chrome Experiments WebGL Globe project for three reason: 1) It could handle 500,000+ location data points in the browser 2) It allowed dead simple mapping of distance in a visually interesting way 3) It was open source

 We use Redis for logging user events which then get pushed to log files.  We’re given the freedom to use whichever language makes us productive so for this task I turned to Python writing a parser for the log events that extracted location, distance of Bump (mapped to the magnitude), and platform information to a JSON file in the following format:

Despite Chrome being the only supported browser for now, it’s a really simple and visually interesting way to display multi-variable data sets. There’s something magical and humbling watching million of people, from all corners of the world, connect, meet, and share their lives.

After all, it’s all about connecting those dots...