My Final Project
NARMADA - Needs and Available Resource Matching Assistant for Disasters and Adversities
Demo Link
https://osm-dm-kgp.github.io/Narmada/
Link to Code
https://github.com/OSM-DM-KGP/Narmada-dlnlp/
How I built it (what's the stack? did I run into issues or discover something new along the way?)
My work was on the backend server. The major services provided by the backend server include classification and categorisation of the tweets in the system. It also provides support for the addition of new information and their automatic categorisation. Facilities have been provided for marking resources once their need is fulfilled or the availability gets exhausted.
The server side uses NodeJS framework and is written in Javascript. Nginx is used as an HTTP server to make the frontend accessible to the public. However, the NLP-related extraction tasks are handled better in Python.
The server partly uses a Flask-based Python backend, a micro web framework.
The Flask server makes API calls to the deep learning classifiers, featuring BERT, which returns the output. The output is further reflected in the frontend. The server sends information requested by the user interface via RESTful API, which supports cached responses on the frontend and enables the system to be scalable, thus allowing more users to use this service.
API endpoints are publicly available, which would allow programmatic access to the server's functionalities (see https://osm-dm-kgp.github.io/Narmada/).
Additional Thoughts / Feelings / Stories
The main objective of our project is to assist and facilitate the coordination of post-disaster relief operations. The work is to be presented at the SocialNLP workshop to be held along with the 2020 Annual Conference of the Association for Computational Linguistics (ACL 2020).
Graduation quote: Tamaso Maa Jyotirgamaya (Lead me from darkness to light)
Top comments (0)