Swipe Style works by presenting you with a series of items that you can swipe left or right on, similar to the popular dating app Tinder. If you swipe right on an item, tags associated with the item are saved in your individual user preferences object. If you swipe left, the item will be dismissed, and the tags will be removed from your user preferences, and you will be presented with the next item. As you swipe through items, Swipe Style uses machine learning algorithms to identify tags associated with the items you like. It then suggests similar items based on these tags so you can discover even more items that match your interests. If you find an item that you like, you can add it to your favourites list, where you can easily access it later. You can also add items to your cart for purchase. One of the unique features of Swipe Style is the way it learns from your preferences over time. As you continue to use the app and swipe on more items, Swipe Style will become better at predicting which items you are likely to be interested in. This means that over time, the suggestions that Swipe Style provides will become even more tailored to your individual preferences.

We used PostgreSQL, Express.js, React Native, Expo, Node.js, Firebase, Git, Jest, Supertest, Axios.

We used PostgreSQL because we needed a relational database known for scalability and reliability. We also chose react native for the front end due to its cross platform capabilities and we used firebase due to it being a popular, accessible user authentication system. All the other tech we chose was based off industry standards.

We faced some challenges with Firebase because most of the documentation was on older versions and so it was hard to find resources on the latest version. We also struggled with getting the algorithm to recommend users the correct clothes based off of there swipe history, however it is much more accurate now. We often faced a bug where the same items were shown multiple times but we have now implemented a recent items table in the database to fix this.