The name of my project is Food Find. Food Find is a chatbot that takes the names of grocery items and searches for stores with those items. The goal of Food Find is to simplify grocery shopping by providing people with a simple way to find many places where they can get a particular item without having to search for that item in multiple places online.
Category Submission: COVID-19 Communications
Demo Link: http://1f3a781d.ngrok.io
Link to Code: https://github.com/YnotCode/Food-Find
How I built it:
I used Node.js to create a server that takes in POST requests from the Twilio Programmable SMS API and responds accordingly. The server also serves static HTML pages and accepts POST requests from those as well. The server integrates the Kroger API and screen scrapes from the Walmart website in order to get links. The server also uses a small database of words that are NOT food items as well as a database of common grocery items to try and interpret natural language. As a beginner, the main challenges I faced were attempting to achieve consistent results with the natural language interpretation and delivering links in a reasonable amount of time.
It was challenging to screen scrape quickly and it was hard to interpret texts without much knowledge on machine learning. The result of my efforts is far from perfect and only integrates Walmart and Kroger, but
I think it can be improved and expanded with more efficient processing and more stores.
A project by Anthony Varkey @ YnotCode on GitHub