Inspiration
Simple university student looking to save a few dollars with a simple picture. Being able to find prices of a certain item with a few clicks can simplify the stream line process of saving money on groceries. Another application of taking a picture of food is being able to look up grocery prices in foreign places. For example if you're on business meeting for a long period and rather then going out to eat, you can save a few dollars on groceries.
What it does
It a web app that uses IBM's Watson's Visual Recognition to help classify images with a primary focus on food. The user will get an option of choosing from the top 3 results from what the visual recognition is able to pick up for a specific item. Using flipp's website to search and parse information about such item. It should display the location, price, store and item for the user to go out and buy.
How I built it
It was used with Angular cli to help make the server using multiple RESTful API calls for the implementation of Watson's Visual Recognition API. There is a lot of focus on JSON parsing because we have to go through the response of the Visual Recognition API and go through the information on Flipp's website.
Challenges I ran into
There are a few major problems we ran into when making our project. The first one was being able to call the Watson's API. It initially was using AngularJS on the front end and Spring on the back end to communicate with each other. Initially we tried to make the RESTful calls through Java but, we ran into some problems and found out they had their own Java Libraries to use. However, we ended up scrapping it because we ran into an error about a mismatch boundary when trying to make calls from our front end to the Spring servers.
The second problem is trying to link the information between the pages. Since we were passing big data across pages we had to figure out a elegant way of passing the information across the pages with what we had. We spent a long time working on this and ran out of time. We have 3 main pages we communicate with and we only managed to get 2 pages working. However, for our first hackathon (besides one of us) we got surprisingly far for such an ambitious product and we are quite happy with our outcome.
Accomplishments that I'm proud of
Paul - I think that I am most proud of from this event is being able to take away everything I have learnt from my teammates and especially the mentors that were able to help us. Being able to see all these mentors and their way of thinking, tackling our problems has been really inspiring and taught me a lot for future hackathons.
This was the first hackathon for most of us and I think we were able to take a lot from this and we were surprised on how close we were to finishing our first hackathon product.
What I learned
I think there is a lot for us to take away because it was our first hackathon. We got to use and learn a lot about new APIs and implementations. Planning and organizing our tasks seemed to work out pretty well and a majority of our work we were able to combine. However, I think our choice of the architecture of our program wasn't to well thought out. We kind of choose whatever, and didn't think to much about the implementations. I think a major thing to take away is when you run into problems brute forcing your way isn't always the best, taking breaks and asking mentors can go a long way then one might think.
Log in or sign up for Devpost to join the conversation.