Building the Map of Twitter

Our next talk is by Mike Barry on visualizing Twitter on a map, a tale of machine learning meets geography meets visualization.

Schedule:

6:30 - 7:15

Mingle and share demos of what you're currently working on. Bring a laptop for some casual show and tell! 

7:15-8:00

The main talk followed by Q&A

-----

We are trying something new! At our last meetup, we lamented the lack of pizza and beer collectively. We've relied on sponsorships so far, but they are unpredictable and we can't always guarantee we'll find them. You know what IS predictable? Hunger. As a result, we are trying a new thing for this meetup - the $5 charge. Your five dollars will get you some pizza and beer (or a non alcoholic beverage). Any excess funds will remain in the pizza and beer fund for future meetups (though our estimates of $5 should cover our costs.)

-----

Talk Abstract:

How can you represent something as rich and complex as the interest space of Twitter in a single eye-span? Recently some data scientists began applying techniques from machine learning to understand this space and Mike built an interactive visualization to browse it.

This talk will cover the simple yet effective machine learning techniques we used and a live demo of the visualization we built which has proved useful across many parts of Twitter. Mike will also cover some libraries and tools we used to interact with hundreds of thousands of data points in the browser at 60fps.

About our Speaker:

Mike is a software engineer at Twitter who leads a team building backend services for data analysis.  He has a passion for data visualization and tinkers with projects in his spare time. Mike also built Visualizing MBTA Data (mbtaviz.github.io) for a graduate course in data visualization at WPI and occasionally gets to do some visualization work at Twitter.