TwitterVis

• 

  Bernie sanders sentiment heat map 2/23

• 

  June 2011 Arab Spring sentiment heat map

Inspiration

The global trend of grassroots movements towards political change

What it does

Provides some heat maps based on twitter sentiments of #arab spring and creates a new heatmap based on tweets from the last week of a given hashtag

How I built it

• Arab Spring

  • python2 script to scrape older tweets (API only goes back one week)
  • write tweets to csv files
  • pass csv files to python 3 script
  • read csv into textblob
  • use tweepy api to lookup location of user using mapquest
  • use textblob to translate and calculate tweet sentiment
  • plot locations and sentiments on matplotlib map

• 1-week map

  • same as above, only pulling tweets directly with tweepy api

• image comparison

  • iterates through arab spring maps and identifies the one most similar to the one-week map
  • identification with open cv

Challenges I ran into

• Data collection for arab spring tweets took much longer than anticipated
• ran into rate limits on tweepy
• bit off too much and made some mistakes later than we could correct them

Accomplishments that I'm proud of

• project from idea to prototype
• used github

What I learned

• learned github with pycharm
• webscraping without apis

What's next for TwitterVis

• improve visualizations
• collect data for more hashtags(climate change, clinton/trump 2016, etc.)
• specialized map comparison algorithm
  • opencv is not necessarily good for this purpose
• create an actual name
• website

Built With

• python