S

Sentiment Analysis Visualization Dashboard

Predict the people

S

Sentiment Analysis Visualization Dashboard

Predict the people

The problem Sentiment Analysis Visualization Dashboard solves

The main objective of this application is to analyze the current tweets related to COVID - 19 and lockdown. By analyzing the tweets, We can understand the sentiment levels (Anger, Joy, Sadness, Disgust, Fear) of the people in the country. Those analyzed data can be represented in various forms of charts and graphs, which will be very useful to understand. We can also predict the future sentiment levels of the people, which will be very useful for the government to understand the response for a particular decision. For example, If lockdown gets extended.

Solution Overview

Data Collection

First of all, We need to collect the current tweets about a particular hashtag. There may be thousands of tweets about a topic, So it may be difficult to collect all those tweets at once. So, We have designed an API which will collect the tweets about a particular topic for every 8 hours. Those collected data will be analyzed and the sentiment levels are obtained by using the IBM Tone Analyzer service. All these data are stored in the database with their time and date of extraction.

Data Visualization

All these analyzed data are visualized in various forms of charts and graphs. The sentiment levels are also visualized in a map with their sentiment range such as positive, negative, and neutral.

Future Sentiment Levels Prediction

We have developed our Machine Learning Model, which predicts the future sentiment levels of the people. To build the model, we have used Scikit learn the library.

Challenges we ran into

We had many challenges while developing this project. One of them was the real-time data collection part. The problem is, If we try to collect the tweets for a particular day, it may be difficult to analyze and the predict the values because the data will be huge. So we made the API to collect the tweets for every 8 hours. So that we can break down the data into smaller parts and analyze them easily

Discussion