Inspiration

We originally wanted to work on something with the Arduino and randomness, but during our planning phases, we found out about Project Oxford at Microsoft. We thought that it would be a cool idea to use this and Twitter to create an automated bot that can analyze people's' emotions in images. Twitter is a very polarizing website, with users having to express themselves in short posts or images. Therefore we think it's the perfect environment for our bot.

What it does

The Feels Bot analyzes the photos of users who follow it and tweets the emotions in the photos presented. We use the API from Microsoft's Project Oxford for the analysis of the images, and Twitter's API for the development of the bot.

Challenges I ran into

We had difficulties in accessing the key from Microsoft in order to have permission to use Oxford. In addition, combining the two APIs was more difficult than we first imagined. We had to figure out how to use Ruby or C# in order to use the Oxford Project API, before realizing that we could use Python.

What's next for Feels Bot

We could further improve Feels Bot in order to detect positive or negativity in text. This could be useful information for companies or notable accounts in order to see the effects of their tweets on consumers.

Share this project:

Updates