TweEmotions

Inspiration

A lot of cool data maps have come out recently, with social media networks such as Snapchat, Instagram, Facebook showing visual data regarding user location and application use. Imagining Twitter to compete on this visual scale was a challenge we thought we would undertake.

What it does

TweEmotions takes real-time tweets from around the world and assess the general feeling of the tweet based on the sentiment analysis functionality of Python's textblob. You can select the color-coded dots in order to view the tweet and the location of the tweeter.

How I built it

Used Tweepy's streaming services in Python to post all real-time tweets and tweeter location into .csv files. Analyzed each tweet's general emotion on a scale of negative, neutral, to positive using textblob analysis. Used ArcGIS platform in order to host data and visualize it.

Challenges I ran into

Understanding the need to convert tweets from UTF-8 to ASCII, incorporating ArcGIS into Python code, visually representing data on ArcGIS' platform in easy way, automating Python script. The limiting number of credits on ArcGIS also proved to be a problem in our data analysis.

Accomplishments that I'm proud of

Completing this task at our first ever hackathon.

What I learned

Python implementation, Tweepy streaming API, API implementation, ArcGIS platform use, textblob sentiment analysis.

What's next for TweEmotions

1. Web application implementation. 2. Using more of ArcGIS' API. 3. Hosting a self-updating program on a server. 4. Moving to 3D models for more pleasing visual display.