Responsive Diary is a web application that takes the user's diary entry (input), locates key words and emotions, and suggest links (to videos, articles, etc) that might be interesting to the user based on how their day went and what happened during their day. Quite like a personal friend after a long day!
How I built it
We implemented it using angularjs, html, and css in the front end, and hapijs and node for the back-end. We processed the user's intentions using the Indico API and found relevant data from the Youtube API and data scraping sites such as CNN and Thought Catalog.
Challenges we ran into
The Youtube Freebase API was deprecated, so we had a huge hurdle to cross in trying to use the very new, very raw WikiData API for our relevant searches. We also had to design a data search model that was feasible to implement in the short amount of time we had.
Accomplishments that I'm proud of
It was the first time I had to implement in less than 24 hours that was up and working, and I've never attempted to finish a project of this scale in those hours. I am very proud of the number of APIs I got working, as well as how I learned how to serve CSS/HTML locally and have it interact with back-end (as I've done mostly back-end in the past).
What I learned
Coffee/red bull is key.
What's next for ResponseDiary
We would like to expand on the model to getting relevant data, perhaps doing simultaneous data scraping of many different website home pages instead of sequentially (right now we only data scrape Thought Catalog), using some sort of parallel computing method. Also to integrate bootstrap in the front-end UI.