Event Lens

Inspiration

We have an interest in machine learning and wanted to try it out on some local Baltimore data.

What it does

It takes event data and finds tweets that are in the same area around the same time and measures the sentiment on the tweets through IBM Watson's API. The app pipes the data from different APIs including Twitter and Watson, and aggregates that data into the front end for the user to visualize the sentiment analysis of the given tweets within a given location.

How we built it

We coded in Clojure and Clojurescript and used Datomic + Datascript for data/query. We used a variety of open source libraries, including:

React
Google MDL
reagent + re-frame + re-thread + re-mdl (cljs abstractions)
httpkit webserver + Sente websockets abstraction

And these APIs:

Google geo
IBM Watson
Twitter Search
Baltimore City Open Data

Challenges we ran into

Time, time, time!

Accomplishments that we're proud of

Bringing all the data together is awesome- really cool to see the genuine sentiment of locals to the event. This app could not have been put together a few years ago- the future is here! This weekend we also invited a member of our Business Development team to hack along with us. So as a whole, the project was not only a fun but also a learning experience for one of our colleagues.

What we learned

Spending the weekend on one project, our team really utilized each other's skill sets to come together and make an awesome web app in a very short amount of time. Our team also came together to teach Allie, a BD team member, about the daily lives of the engineering team at our company. We really grew as a team! Thanks for having us!