/Important-Twitter-Nodes

Primary LanguagePython

Notable Nodes

This project is designed to illustrate the spread of information about a particular topic on twitter. Currently the app can be run in interactive and terminal mode


Demo Video


Running the entire Project in terminal mode

Install dependencies

#Install required dependencies then run app
$ pip3 install -r requirements.txt

Running the entire Project in interactive mode

#Install required dependencies then run app
$ pip3 install -r requirements.txt
$ python3 app.py

Run in default terminal mode

$ python3 app.py terminal

Run the offline module

# tweets wont be downloaded
$ python3 app.py terminal offline

NOTE: To bypass the 50 tweet limit

    input a search term
    $ example%
    #add the % character to bypass the 50 tweet limit

Specific File instructions

Run processTweets.py to collect tweets that contain specified term in the terminal. This module also contains the twitterSpider class which can be imported into the ui module

Please create a .env file and fill out the appropriate environment variables

Tweets will be saved into a EACOPV3.csv file

As of now, most of the graph experimentation with networkX is happening in graphing.py. In interactive mode, we can see the different subgraphs however, it only defaultlu works with the House of the Dragon dataset. An error I encountered is that I needed to reconfigrue the way I draw the graphs before pickling them.

NOTE:

The House of the Dragon dataset is currently the largest default dataset with 50,000 tweets.

The graph drawn in GUI / Interactive mode is static when to compared to the one rendered by the terminal mode.

The Project Proposal is williamProposal15112.pdf

data.csv and EACOPV3.csv are sample files that contain collected twitter data.

TO DO's:

  • Add EC2 script that will be used to analyze the tweets

  • Unpickle the graphs when the buttons in the UI are clicked

  • Improve the interactive mode with PyInquirer

  • Find a way to make the graphs more interactive when using the GUI

  • Import twitter spider into the frontend

  • Impelment regex to extract twitter data for a tweet from a user inputed URL

  • Port networkX code into main UI.

  • Implement page rank with networkX

  • Find way to display these results on the screen