A Python application that uses the Flask web framework and Bokeh interactive visualisations, deployed to Heroku. What is it?
With one click, you can see the number of times marijuana-related words are currently being tweeted in a region of your choice! What does it do?
See how much different regions tweet about a topic of your interest. What coud it ultimately do?
Example 1: Do people tweet more about marijuana in states where its use is legal?
Example 2: Do people tweet more about your product in places where you focused your marketing campaign?
1. Get current tweets from Twitter API which mention marijuana. How does it do it?
2. Tokenize them.
3. Count the number of times marijuana-related words are used.
4. Return the number of times marijuana-related words are used in an interactive graph.
1. At the moment, the user can choose from West USA, East USA and Canada. The app counts marijuana-related words in the first 3 tweets in real time. What is the current state of the app?
2. I'm working on analyzing more tweets. However, analyzing the first e.g. 10 tweets per region in real time takes me over the Heroku HTTP 30 second timeout limit.
3. My next step is for the user to be able to choose more regions at once and see their comparison.
4. My ultimate step is for the user to be able to choose any topic of interest and compare the volume of tweets about the topic across different regions.