It goes without saying, 2020 was extremely challenging for us to get together and connect with other people. Dates fell through due to the difficulty to foresee things that would cancel a plan.
We were literally inspired by our frustrations and wanted to create something that would gently remind people of planning mishaps and also propose alternative ideas!
What it does
Catch is a mobile app that helps users plan dates better by checking logistical problems. We’re only human, and sometimes life gets so busy we don’t check everything. Catch has got you covered, by checking your date for you.
It also suggests related date ideas, giving you plan b if things fall through.
How we built it
The designers first sketched out the flow and screens, before moving it to Figma. Using sketches, we agreed on the general direction. Afterwards, most of the design work was completed on Figma. Images and icons were made with Illustrator and some free image libraries (splash, undraw).
For the coding side: Android studio was the main software used. The developers (Eunice and Zoe) together implemented the back and front-end, created a sample dataset, where Zoe developed and trained the intelligent recommendation system in Tensorflow. Eunice developed the front-end in Android Studio. We also deployed the recommendation model and connected with the mobile app.
Challenges we ran into
Since half our team was remote, communication was difficult at times. Our schedule had conflict and sometimes we were unable to all be online at the same time.
Time was another constraint. We built a dataset with nearly 750 samples, but due to time limitation, the training labels are not full fit into the real life situations. If there's more time spent on refining dataset, the model accuray will be much improved.
Accomplishments that we're proud of
We're proud of how fun this idea is, and how we're able to put it together in such a short timeline. As a half-design half-dev team, we put together something that played to our strengths.
What we learned
We learned that there's a lot of ways to date. But also that it's hard to plan and meet and date properly. Sometimes, defining these "inputs" and dating categories are difficult and needs a second look.
What's next for Catch
Catch can extend to plan different types of events, employ a subscription model, and improve its algorithm to create an exploration feature. The possibilities are endless.