CoffeeRun

Inspiration

I'm interested in NLP, so I wanted to work with one of the provided APIs that were related. Nuance had such a toolkit and they talked about how you could use it to order food. While talking with a friend, we realized that it was always difficult to keep track of coffee orders, so it would be possible to train a model to recognize the different parts of a coffee order.

What it does

For everyone who struggles to understand what an "iced coffee with two pumps of caramel and one scoop of macha powder" is or to remember that their friend asked for "a venti pumpkin spice latte with no whipped cream and skimmed milk", this app will help them remember all of their friends' orders, and it tags the important parts of each order. So, it recognizes orders of coffee and tags the different concepts (drink, size, milk-type, amount of sugar, etc.)

How I built it

• ran through an iOS tutorial to make a basic list app
• developed a model using Nuance's toolkit to have it recognize the different orders
• combined list application with sample app provided by Nuance

Challenges I ran into

• coming up with a way to break down coffee order
• making iOS work, debugging errors in code (!!)

Accomplishments that I'm proud of

-having a functioning mobile app

What I learned

• learning Objective-C
• how to train and test a model

What's next for CoffeeRun

• Fixing minor bugs on app
• Calculating prices for each order: parsing order info and comparing against a menu
• Request for payments from friends

Built With

• nuance's-nlu-&-asr-toolkit
• objective-c