/election-prediction-system

R-Based Analysis System to generate election trends based on twitter data.

Primary LanguageR

Election Prediction using Twitter Data by means of vector modeling.

This is R project that is done to analyse data from twitter to prediction popularity of candidates for elections.

What it does

  • Extract Raw Data from Twitter and Store it into Mongo DB.
  • Process the stored data and filter the data based on following parameter
    • Location of phenomenon
    • Search Terms related to election
    • Number of tweets
    • Time window of the phenomenon
    • User Profile characteristics
  • Calculates the following parameters
    • Relative Frequency - Ratio of total tweets per day to no. of tweets of particular party.
    • Positive Negative Ratio - Ratio of positive no. of tweets to that of negative no. of tweets.

Steps to execute the project

  • Download / Clone the repository
  • Install Mongo DB
  • Start Mongo Database in localhost
  • Run command mongod.exe --dbpath 'path_to_database_folder' Rscript main.R
  • Plots will be generated in a PDF file.

Modules contained in the project

  • function.R -- function module containing the main functions to be used in the project
  • environment.R -- contains the global parameters to be used
  • packages.R -- contains the list of libraries / R packages to be used in the project
  • main.R -- main module where the project runs.
  • mine.R -- module that must be executed to mine data from twitter API as needed.

Database Configurations

  • Start Mongo DB server and that's it. You database is ready to be used.