Skip to content

Home

Jump to bottom
Anubhav Gupta edited this page Nov 25, 2022 · 4 revisions

Welcome to the disasteropedia wiki!

What is it ? 🤔💭

This is a website for monitoring the disaster situation in India through Social Media posts. It is mainly made for the government agencies in India, like NDRF (National Disaster Relief Force)

Any post which is shared on Social Media (currently supports Twitter only), which is about a disaster, gets shown here on the website, along with the media of that disaster, in real time (LIVE). It finds the important data from the tweets, like the location and type of disaster. All this data is shown in a user friendly & organized manner, to make it easier for the authorities to monitor the situation.

Technologies Used 🖥️

  1. ReactJS
  2. Django Rest Framework
  3. Twitter API v2

Architecture 📊

Below is a high level architecture diagram for the disasteropedia app, to help you understand how it works:

disasteropedia-architecture

Abstract 📄

We check for disaster specific tweets from Twitter in real time, and show it on the web application along with some analytics of it.
However, obtaining tweets is not an easy task as we need to filter out disaster specific tweets of India from 1000's of tweets which are posted every second on Twitter (which are not relevant to us).
This tweet filtration process is divided into 3 levels:

  1. Tweet Filtration Level #1:
    This is the very first filtration level which filters out the majority of tweets. The disasteropedia-twitter repository handles this. It obtains tweets from Twitter in real time via Twitter API and filter's it out based on specific keywords. It then passes the filtered tweets to the second filtration level. You can refer to it's docs for more details.

  2. Tweet Filtration Level #2:
    This filtration level is where Natural Language Processing (NLP) comes into picture. NLP is applied on the received tweets and important data is extracted from it, like the type of disaster and location of disaster. Accordingly, the tweet is discarded if it's not a relevant tweet about a disaster. Otherwise, if it's a relevant tweet, then that tweet is passed to the next filtration level.

  3. Tweet Filtration Level #3:
    This is the final level of filtration. Here, the location of disaster is analyzed. We find the state of disaster and verify that the location is actually a place in India, with the help of Wikipedia API.
    Once verified, we finally save that tweet to our database.

Clone this wiki locally