Food Menu Keyword Indexer for Better Recommendations

• 

  FMKIFBR

Inspiration

Asian restaurants have massive, unwieldy menus that take hours to comb through. How is this so inefficient? What should I eat?

Why do Yelp, Google Reviews, and all other food review places include poor ratings because of service? What if I only care about how the food tastes (since I'm already in the restaurant)?

What it does

Scan menu and identify word chunks in real-time. User can select a chunk and then select specific keywords that corresponded to the dish Based on the keywords, user can rate the dish/food-stuff

--- ML implementation not included in submission ---

Recommend dish to user

How I built it

Android Studio Google OCR resources

Challenges I ran into

ML implementation that is not shown in this codebase was particularly difficult to work with

Accomplishments that I'm proud of

Develop Android application

What I learned

One of first times learning complex Android, so turned to many tutorials and read through a lot of what goes into Google's OCR. Also tried many different Google Cloud products to understand how each could fit into project later.

What's next for Food Menu Keyword Indexer for Better Recommendations

Train better ML model in order to have accurate recommendations like Amazon's "Customers who bought this item also..." or "Recommended based on your (eating habits)" Fully flush out back-end server-side work, which was not entirely functional and therefore not included

Built With

• android
• google-cloud
• java
• ocr