PROJECT DESCRIPTION
The adoption of social media and its use for the widespread dissemination of political
information and sentiment is so remarkable that it has impacted traditional media. Nowadays,
Twitter is a convenient tool for journalists in search of quotes from prominent news sources,
e.g., politicians, as they can add direct quotes to stories without having the source in front of a
microphone or camera.
Following this, we use a sentiment calculator, trend analyzer, and influence calculator to deduce
the opinion about the proposed policy by the government. For example, we use keywords like
GST in sentiment calculator to find the overall sentiment and opinion of people on the topic.
We represent the result in the form of charts and graphs. Thus, the result can be used by the
government and its people to know about the reaction of the public concerning the implemented
decision. This way, the government can make necessary amendments to the policy if needed.
Thus, public opinion can be included in making various policy decisions.
METHODOLOGY AND MODULES
The project is a framework for data analysis and network capturing of users. It uses VADER
semantic analysis for semantic analysis of the tweets, to get the opinion based on keywords and
scores. VADER (Valence Aware Dictionary and Sentiment Reasoner), it is a rule based
semantic and sentiment analyser which is specially tuned to analyses social media data.
The
VADER scores each lexicon and sums the score and it is normalized to get a compound score.
Streamlit a framework for seamless integration of data and the scripts for its analysis.
Tweet analyser on a particular political keyword, charts used to represent data the sentiment
and the overall count of the words. This used VADER. Individual profile of the user giving
general information of the user and also calculating the influential rate. It is determined by the
parameters such as total amount of twitter user in India, total followers of the particular user
and the finding his/her first circle of friends (about top 25) their followers and calculating a
ration to find the influential rate. A database for politicians’ data, division, party, state and
generic information.
Analysing the social circle for a user. Finding and generating the social
connections graph for a user and is also done for a particular keyword (E.g., BJp4india, gst).
26
Influence rate is calculated based on the user’s followers and the number of followers his top
friends/followers have, by using a weighted model it is calculated.
Influence Rate=(0.75*(Number of followers/total users in India))+(0.25*(sum of followers
of his followers/total users))
• INFLUENTIAL RATE
Using Twitter API, we calculate the influential rate of the User.
On the basis of a keyword, hashtag, user a network graph is generated for the
user/keyword. For the user, the first-degree friends are mapped. For keywords, the
tweets which have that keyword are extracted and the users and the people mentioned
in that tweet are mapped.
• SEMANTIC ANALYSIS
The project is a framework for data analysis and network capturing of users. First, the
tweets are extracted, for extracting the tweets a package that scrapes the tweets has been
used. After extraction is done using the VADER analyzer the polarity of tweets was
calculated, that is, it uses VADER semantic analysis for semantic analysis of the tweets,
to get the opinion based on keywords and scores. These scores corresponding to polarity
saved in the final data file. Now when a user enters a keyword such as “BJP” the tweets
with that keyword and the polarity of them appear.
• TRENDING LIST
Using Twitter API and tweepy based on the yahoo geolocation ID we get the trends in
a particular region (India) and display them. Libraries used for all these are tweepy and
streamlit
.
• SUGGESTIONS
The government/user can search/scrape through twitter regarding suggestions for a policy and
hence get the public view
If you liked/found this project helpful , please support the authors by commenting 🙂