educATE

Want healthy, affordable recipes? Don't know what to make? educAte's got your back! With a simple food image input, we'll recommend you recipes based on past food choices, and a grocery list to boot!

Comment

Inspiration

Us being constantly hungry students, a food-related app was on the top of our list! As university students, we are constantly faced with the challenge of juggling academics and domestic life. As a result, many of us often resort to eating out, sometimes cultivating unhealthy eating habits. We therefore wanted to create an app that allows users to find in their kitchen the same eating experience they find in restaurants, while streamlining the process of finding balanced recipes and good ingredients.

What it does

As indecisive users, we aimed for minimal user input. Thus, we require only images of food as input from the user, in addition to dietary requirements. We generate the location based on the image metadata, and look up the closest restaurant to the geolocation tag. We then identify the cuisine(s) of the restaurant and use it to generate a list of 3 recipes for the user to choose from.

We also webscrape Amazon to find a realistic baseline price for the groceries required, and provide the user with the prices of the ingredients as well. This is all displayed in a user-friendly, simple-to-use web application.

How we built it

We've used Python and Restful APIs to generate location metadata based on an image input and generate a list of recipe recommendations with a set of search parameters, with the help of Spoonacular API. We've used Elixir and Phoenix to webscrape Amazon for groceries. Our frontend is build with ReactJS and TailwindCSS.

Challenges we ran into

A lot of bugs. In Python mostly. And Elixir. And integrating our parts together because that's when most things break. As EIE students used to low level and strongly typed languages, we were sometimes caught unaware by the dynamic typing and runtime errors in Python.

Accomplishments that we're proud of

That we managed to get a minimum viable working product!

What we learned

Debugging is hard. Was also a very rewarding experience!

What's next for educATE

  • Computer vision/machine learning model to detect details of and tag food pictures.
  • More selective filtering based on food image tags and nutritional value of recipes.
  • More recommendations based on user choice feedback and past user interactions.
  • User rating system and inter-user interaction.

Built With

Share this project:

Updates