Inspiration
We were inspired by some terrible apartment shopping experiences in the past. There are too many apartments and too many factors to consider with all of them, as well as no good way to go through them all.
What it does
Dom changes that by introducing only one apartment at a time to users, learning what they think of it, and showing the next one that Dom predicts the user will like.
How we built it
We started by building an android app that allows for liking or disliking of apartment listings that are shown, then added the Watson functionality to sort the apartments by whether the user will like them. Then, we built up the web app. from that.
Challenges we ran into
As we learned more about Watson we changed the architecture of the application multiple times. Getting everyone on the same page is difficult, especially with rapid sweeping changes that discards hours worth of work.
Accomplishments that we're proud of
We are proud we got everything working together, because that is what we ran into issues with throughout the hackathon. We were also impressed with the end user's experience, that was the most important factor in this project.
What we learned
Big changes are normal for hackathons, but it is important to change plans correctly. Also none of us had Watson experience, and only half of us had any Android experience.
What's next for Dom
The next big step is getting real data, such as from the Zillow API. After that, we add more data points (number of bathrooms, walkscore, etc) so Dom can learn more about the user's preferences. We all believe in the project's potential!
Log in or sign up for Devpost to join the conversation.