FitCast

• 

  Icon

• 

  Image processing

Inspiration

As a group of Montréal natives or long-time residents, we have often experienced the difficulty of styling outfits to fit harsh weather conditions. To make finding the perfect fit easier, even in the toughest weather, we decided to design an application that would store your wardrobe, give you weather-relevant recommendations, and allow you to quickly mix-and-match styles.

What it does

The app does exactly what it says on the tin: it provides outfit inspiration according to the local forecast—a FitCast. Based on the temperature and weather, it provides "styles" that determine the type of clothes should wear to feel comfortable, but allows you to mix-and-match what clothes you want to wear within those styles. Multiple styles can be recommended for any given situation.

How we built it

The app leverages the OpenWeatherMap API, two Machine Learning models to classify and isolate clothing images, and a custom backend to provide the best experience to the user. The backend API is Go, the frontend in React Native, and the machine learning side is done in Python.

Challenges we ran into

Some of the technology was new to us—for example React Native—so we spent a lot of time working on building our skills there. We also had difficulty connecting our Python and Golang backends, in addition to difficulty properly sending many images over an API connection. Further, we had difficultly making the AI models predict well.

Accomplishments that we're proud of

We're very proud of all the challenges we've overcome, and how much progress we made in such little time. The ability to isolate clothing images from an API and run an app on a mobile device are both major achievements for us.

What we learned

We learned a lot about deploying apps to mobile, which is a significantly different experience from the web development we typically do.

What's next for FitCast

In the future we're hoping to improve the UI of the application, improve the processing speed by upgrading the remote server's RAM, and fully flesh out the wardrobe feature that we might not have time to finish.

Built With

• go
• python
• react-native
• typescript