harisankar01 / Twitter-Analysis-and-First-Response-Application

A application which is used to monitor twitter for incidents and complaints that are related to the police department along information extraction about a suspect from twitter.

Twitter Analysis and First Response Application

• The application is built for the police department to check for incidents from Twitter.
• This application provides real-time Twitter analysis about incidents in a state and send email to the respective police department.
• Analysis of tweets to find if it is related to a complaint or police incident
• Collect information about a suspect from twitter and get the timeline of a tweet posted by the suspect
• Provide priority to tweets and respond to tweets directly from the application.
• Displays the tweets tagged or mentioned to the respective state's police department.
• In-built chat feature which can be used to chat between users.
• A login system based on the state and city in India(a representation of important police stations in a city).
• Get the complaints made on Twitter to different departments of police
• Populating Twitter data automatically using azure functions and time-based triggers.
• Tagging of geo hotspots of crime using Twitter data(in development).

Tech Stack

• React with Next JS(Frontend)
• Backend: Node JS (API endpoints and API routes in next js) -> Primarily used for MongoDB and posting of tweets
• Twitter tweet extracter: Python Flask API => Hosted in Azure Web app --> Extract tweets based on paramters
• Backend(Tweet analyser): Python Flask API => Hosted in Azure web app --> Analysis

Note: The tweet analyzer link, tweet extractor link, MongoDB link, and database name must be provided in the .env.local file

Local Development

• Clone the repository

git clone "https://github.com/harisankar01/Twitter-Analysis-and-First-Response-Appliaction.git"

• Move into the cloned folder

cd Twitter-Analysis-and-First-Response-Application

• Install all the packages

npm install

Links

• Create and add the following links in the .env.local file
  • Link to the Twitter extractor backend API(either deployed or running locally) -> twitter-extractor-backend
  • Link to the Twitter analyser backend API(either deployed or running locally) -> Twitter-analyser-backend
  • MongoDB cluster running locally or remotely. A simple option will be to use the MongoDB-Atlas

Note: Visit the repos mentioned above separately to know how to deploy the backend or develop it.

You can refer to the sample.env file for the structure to be used in .env.local.

To run the server, use the following command,

npm run dev

Images

The images have been uploaded to Cloudinary and the link to those uploaded images is present in the MongoDB database.

Azure Development

Azure functions have been used for the following functions,

• Populate the MongoDB database with new data extracted by the Twitter extractor API
• Populate the database with processed data
• Send an email to the user about the incidents A time-based trigger is used to populate the databases.

Please check the image for the structure of the pipeline,

The websites are deployed with azure web app directly from visual studio using the Azure Cloud extension of visual studio.

The configuration of the database used is: Cosmos DB with MongoDB API and is present in the azure cloud Mumbai region.

ScreenShots

The user can log in by selecting the state and the city in India.

The home page loads the tweets that were tagged or mentioned to the police of that city.

The complaints page shows the complaints given to the people of the state

The Chat is featured as a chat room for each tweet

The timeline of tweets posted by suspected people can be viewed in the digital profile section

When a tweet in the timeline module is clicked, it will lead to the page featuring the basic details of the suspect

Addition Remarks

Please feel free to ask anything about the project.