Seeing mass amounts of data in the world untouched and ready to be analyzed and visualized, I wanted to make an app that showcases the power in letting people see the inferences that come from data, and the benefits that can be built up on top of that this data analysis.
What it does
Pull tweets, analyze them using IBM's Alchemy API to analyze the emotion of the tweets, and then plotting them on a map, allowing for inferences to be made such as: where the happiest places are in the world, or in a city. We can also use other analyzation using Alchemy to gain even more inferences from this data, and allowing for even more applications.
How we built it
Twitter tweet-getting module, which sends structured lists of tweets to the next module, which is the AlchemyAPI module. This module processes the tweets for emotion, and then they are passed off to the front end via JSON. We then plot these data points beautifully using the google maps api, with a very easy to use, simple interface.
Challenges we ran into
The twitter API was incredibly difficult to deal with, it was very very inconsistent in it's querying, which made the rest of our development difficult because the twitter is the only source of data that we have, which powers our data-driven app. Alchemy API needed the text to be cleaned in a way, such that we couldn't just push the tweets into the api, but rather we had to clean large amounts of data very tediously. As far as the front end goes, we had troubles, being beginner developers, with doing things like plotting multiple data points in different colors on a map. We also decided to use AngularJS for part of the front end, which proved to be much more than a minor headache...
Accomplishments that we're proud of
We are proud that we were able to design and develop modules that were semi-efficient in a very short period of time. We are proud of the idea, and the scope of the project, and the potential that is contained within this idea.
What we learned
We learned how to learn new technologies very quickly, and how to implement them to get a final product. From AngularJS, to AlchemyAPI, to amadeusAPI, the frustrating twitter API, and more. There was no shortage of learning to be done.
What's next for Em.io
This app idea holds a lot of potential and in the future, we want to design and implement better features, in addition to making the app more efficient, and easier to use.