The inspiration for this project came when we were playing around with the ClarifAI API. Convolutional Neural Nets were interesting and it was fun putting various pictures of ourselves and our friends and seeing what ClarifAI popped out. We saw some really uplifting tags among the other words, and wanted to emphasize these. Our team was inspired by MLH's hack against harassment initiative and wanted to write code that could improve how people saw themselves.
What it does
Compliment Machine takes in an uploaded image and uses the ClarifAI API to obtain tags relevant to the image. It then uses the Natural Language Toolkit to find the most positive of those tags and displays them for the user.
How we built it
We built the backend with "beego" a web framework designed for golang. The frontend was standard HTML5, CSS3, and JS.
Challenges we ran into
Our whole team consisted of people who were most comfortable with backend development. As a result none of us had much experience building attractive web pages on the front. Nonetheless, we were able to overcome this with a lot of time spent looking at examples and documentation.
Accomplishments that we're proud of
Two-thirds of our team had never used go before for server-side development. We are proud of being able to pick up the new language well enough to have contributed to the code.
What we learned
Besides improving our frontend skills and learning a new programming language, we learned how to build a team that worked more efficiently by dividing up team work in a way that people weren't waiting for each other to finish.
What's next for ComplimentMachine
It will definitely be helpful to share ComplimentMachine with people so that they can help themselves. We can also look into work on making corresponding mobile apps that make the functionality more accessible.