A Web Portal for the purpose of helping farmers in crop planning and crop handling built as a part of Course project CS305. The portal gives information on Crop Handling techniques (water requirements, soil type requirement, etc) as well as recommended crops based on attributes such as location. The portal also provides a discussion space for farmers and experts in the related fields.
The queries are answered through Knowledge Models Graphs for each crop extracted from unstructured agricultural text. The co-reference resolution and relation extraction is done through Stanford Core NLP and Snips NLU models for intent and slots identification are fine tuned on agricultural data.
Web Server on NGINX
Django framework for backend design
Databases: MongoDB, MySQL and Neo4j (for graphs)
Tornado based Runtime Push Server
Website Design in HTML/CSS
More information on design of the project in the Design Document: https://github.com/CS305-Group20/FarmerAdvisory/blob/master/Design%20Document%20CS305_Group20.pdf
More demo images/screenshots here: https://github.com/CS305-Group20/FarmerAdvisory/tree/master/demo-images