Make medical centers more accessible to those who need them most. Even though everybody needs to be able to find a hospital in a city with high 911 calling rate, some need it more regularly than others. Additionally, in certain areas within the city where 911 calls are particularly high, it's easy for hospitals to overflow and for people to be left untreated.

What it does

This web app takes in a csv file with each row being a 911 call, and having (at the very least) columns that say 'lat','lng', and 'addr'. These columns should give the latitude, longitude, and address of the incident. The web app also takes the amount of Emergency Medical Centers available and appropriately distributes these Centers in a way that areas with the most 911 calls can easily get help.

How I built it

I built it using a flask python micro-framework and web programming languages. I used python to build the server's framework and to perform machine learning analysis and clustering of regions within the city. The machine learning analysis was done using sci kit-learn. I used web programming languages such as JavaScript, AJAX, JQuery, HTML, and CSS to create the front-end with a clean design. Finally, I used MapBox APIs to plot and visualize the centers' locations.

Challenges I ran into

Using the MapBox API was tough. Since there were so many more 911 calls than centers, I had to figure out how to appropriately zoom in so that the centers are visible and it is clear that the algorithm worked. Analyzing the dataset was also quite difficult given that there were almost 300,000 rows of data that I was analyzing.

Accomplishments that I'm proud of

I'm proud of combining Materialize with MapBox to create such a clean design and a user-friendly interface. Successfully clustering the data was also a big accomplishment because I had never worked with the clustering algorithm before.

What I learned

I learned more about Sci-Kit Learn and MapBox APIs. I've never worked with these much before but I'm glad that I had the opportunity to solve a real-world problem with them.

What's next for Safety First

With more computational resources behind the web app, I can analyze larger datasets and cluster even more effectively. This will ensure maximum accessibility to hospitals in unsafe areas. Given that I was running this on a computer with no GPUs and on one core, the processing power was quite limited. On a bigger system with GPUs and multiprocessing possibilities, Safety First's recommendations will be of significantly more impact.

Share this project: