Aksh [Backend]

Making Data Driven Decisions from Geo-Spatial Data

Won the 4th Prize at Smart India Hackathon 2018 (ISRO).

Android App Source Code: https://github.com/littlewonder/aksh-app

This is the backend of Speaking Geo-GP assistant.

Following is our tech-stack:

• Raw Dataset which includes satellite imagery, metadata and some demographic data provided by ISRO and Census of 2011.

• MongoDB as database since it provides flexible query mechanism for geospatial data.

• NodeJS for backend, making it asynchronous, scalable, and agile. We use ExpressJS to implement a web API which communicates with the Android application.

• NLP library for analyzing command and intent from text.

• Translation library for translating between regional languages.

---

Below is the description of the files and directory structure

• index.js : The entry point of the express app.

• queries.js : This file houses all required methods for geographical queries.

• queryMapping.js : This file maps intent obtained from natural language processing to respective database query.

• populate-database.js : This script reads given data from ISRO and populates MongoDB database to optimize queries.