Based on your liked pictures on Instagram, InstaiEtsy recommends products for you on Etsy!
Using a Flask application, we built a Python backend that authenticates an Instagram user, obtains their recently-liked pictures, and runs them through the Clarifai API. Using the image recognition tags generated, we use a custom machine learning algorithm to recommend products to you from Etsy listings.
We ran into some problems, including properly handling a list of Instagram pictures with the Clarifai API, and how to then use these results with the algorithm. And with the machine learning algorithm itself, a lot of work was needed to get the expected output on training examples, and then make this solution efficient.
We are proud of the final product, specifically how we managed a complex project with many different pieces to work fluidly. And additionally, that we have it all up and running on a public website that looks nice.
We learned a lot about more advanced backend skills, such as having multiple API's working together and having a clean workflow of several HTTP requests and methods interacting seamlessly throughout the life cycle of the running application. Also, the art of collaborating efficiently, as well as debugging and testing the service.
We would like to continue to work on the machine learning algorithm, to enable the backend to make progressively better product recommendations to users. We would like to fine-tune the website, and launch it on a domain.