[Hackathon #10] - Free Your Mind, with Code Nation and EMIS Health

This is the first hackathon I will be attending after becoming a co-founder of INEVITABLE. It was a great opportunity to discuss and develop an idea prior to the event, which is not something I usually get to do.

Initial Idea

My initial idea was to create a platform for monitoring users' social media feeds to predict patterns and perform sentiment analysis. Given the time constraints of the one-day hackathon, it seemed like a reasonable project, even if I ended up working alone.

The idea expanded when my co-founder, Ben, pointed out that one of the signs of suicide could be detected through this platform. According to a clinical article on Recognizing Suicidal Behavior, sudden calmness after a period of depression or moodiness can indicate a person's decision to end their life.

Ben also highlighted two other factors that could predict issues: water usage (how much a person drinks and showers/bathes) and electricity usage (indicating a person's circadian rhythm). Disruptions in these routines can be strong indicators of problems.

This led me to envision a platform that leverages smart meters and social media to help caregivers detect issues and intervene at different stages before they escalate.

The last discussion before the hackathon was about naming this platform. Ben came up with the idea of "Caroline," which is an abbreviation of "Carer Online."

The Hackathon Day

After some pleasantries and pastries, it was time for team building. Many participants came with pre-formed teams and had assigned tables. Those of us who came alone were initially grouped together into one large team of nine. However, I felt it would be more manageable to split into smaller teams.

Unfortunately, the organizers then asked us to refine our ideas and share individual pitches within our groups. With nine people at the table, there were multiple ideas and a lot of "this could be part of that" discussions.

To address this, I suggested that we either focus on a particular piece of technology, like an Echo Dot, and find a problem to solve, or we home in on a specific issue and develop a solution around it. I reiterated my idea, which left me with one other developer on my team, Kevin Thomas, who had experience with the Watson API.

Development

Once we formed our team of two, we divided the tasks. I took on the front-end development to create a visually appealing user interface. Kevin built the necessary endpoints for me to retrieve data from Twitter and perform additional analysis on the text.

After some time, I had a user interface that I was satisfied with, and I was successfully pulling Twitter data through a NodeJS server hosted on Heroku, using the Twit package. As we approached the pizza break, I decided to take a short coffee break before continuing.

While having coffee, Kevin and I discussed the possibility of integrating a Watson chatbot that would send direct messages on Twitter or other connected social networks when it detected "sudden calmness." This intervention mechanism would be especially useful during late-night hours when issues may arise while people are sleeping.

Spot Prize

Shortly after pizza time, as I was reviewing the Swagger documentation Kevin had prepared, a spot prize was announced. This prize was given at a random time for a particular reason.

To our surprise, it turned out that the spot prize was awarded to our team for our pizza handling techniques. We received a pair of Raspberry Pi 3 Model B+ as our reward.

Last Hour of Development

With only one hour left, I focused on integrating the API endpoints Kevin had created into the front-end using jQuery Ajax calls. Initially, I encountered an issue due to an oversight on my part.

Fortunately, in the last 5-10 minutes, I managed to retrieve the data in JSON format, although it wasn't formatted as intended for the demo. I made a mental note to format it properly later.

Demo Time

Our development time came to an end, and it was time for the demos. Since there were several teams, it took a while, and we nearly exceeded the allocated time. Unfortunately, I had been so engrossed in development that I didn't have a proper slideshow prepared, covering specific use cases, the business case, and relevant statistics. It was definitely not one of my best pitches.

Regrettably, we didn't rank in the top three, but we had fun throughout the event, especially after receiving the spot prize. I made modifications to the front-end demo, ensuring the additional information from Kevin was formatted as originally intended. I also made the same repository public on GitHub.