Twitter Tweets analysis using Spatial Algorithms

This project investigates how to organize and classifya large collection of real time Twitter data, working with a dataset of almost 3 Lakhs tweets collected directly from Twitter.The approach described here combines spatial analysis of thelocation of the tweet with content/sentiment analysis of the textand hash-tags associated with the same tweet. We then lookinto those specific regions to try and identify the most popular Political Party and its influence on the regions and comparedifferences between each city, to provide a better plan for the Political Party Campaign. We expect our results to vary basedon where people are tweeting from and which Political Partyhas more impact in the city.

Tech

We used a number of libraries and API.

• Python
• Tweepy API to fetch Tweets
• googletrans to convert HINGLISH word to ENGLISH word
• csv to handle csv file and data
• Textblob to find the senitment of the tweets

Installation

Install Tweepy

$ pip install tweepy

Install googletrans

$ pip install googletrans

Install textblob

$ pip install -U textblob
$ python -m textblob.download_corpora

Development

After directing to the project folder run the python file. Collect the tweets from Twitter

$ python tweets_final.py

Clean the tweets, convert Hinglish tweets to English, and find sentiment of each tweet.

$ python sentiment.py

CLuster tweets on the basis of different location in the previous dataset and count total number of positive, negative and neutral sentiments of that location.

$ python counting.py

Plot the countbjpcity.csv and countcongresscity.csv using Qgis and its library MMQGIS and OpenLayerMap.