It all began three weeks ago when Charlie, Gerry, and Gene convinced Hector to join them in the hackathon.

The idea was not born until late in the wee hours of the night when everyone was fed up with Charlie's "one way photo encryption" idea and instead wanted something more useful. Thus we decided visualize data over time and settled on Barack Obama as a viable candidate for our experimentation.

We used Microsoft's Bing Search to find the images, along with their Emotions API via Microsoft Cognitive Services. We used python as our backend for the number crunching. Then we fed all our pickled info over to a csv file in order to add it to a Google Spreadsheet. From there, we found it easy to use Google's Google Chart API to make different plots of our data.

However, our team did not underestimated the importance of user experience. Charlie and Gene worked together to display the data in the most easy-to-understand and the most easily-accessible way. To do this, they used Bootstrap to format the charts and the website, and deployed the website using Microsoft Azure.

We tried to compare Obama's emotions to different values such as S&P, and approval rate. Unfortunately, no matter how much we tried, we didn't see any correlations between the data (this could be due to Emotion API's inclination to classify a face as happy or neutral more than anything else). Of course, if there is one trend we managed to see, it's that Obama's happiness has dropped sharply this past month and there has been a slight rise in his neutrality, contempt and anger.

Overall this was a great project that we all enjoyed working on and we hope you all can enjoy looking at the data we've gotten for you.

Built With

Share this project: