Since the Coronavirus has such a big effect on so many lives and we are passionate about data science, this app was the natural combination of the two.
What it does
The Coronavirus tracker is an interactive map that visualizes the possible spread of the Coronavirus. The infection rate and the cure rate were modeled to predict what the chance of the virus spreading was. Then, the data is visualized on a vector map. The prediction was not quite finished on the application itself.
How we built it
We started in Logger pro modeling the infection and cure rates of the Coronavirus and did lots of research. We looked at map APIs in order to determine the best possible way to predict how it would spread. We made a frontend in node to show this data.
Challenges we ran into
The data could be modeled in a quadratic or an exponential due to suspicious and maybe inaccurate data from the chinese government, so finding accurate info was hard. We overcame this by searching together and finding a trustworthy source with more complete data. We also found that some of the files for maps were very big, like 680 gb, and overcame that by discarding irrelevant data and doing data analysis. Also the map APIs were very difficult to use, but we overcame that by switching the map logic from server side to client side.
Accomplishments that we're proud of
Finishing our first hackathon!
What we learned
To be more cautious in terms of potential setbacks, as there were a few bottlenecks that took out the rest of the team by everyone waiting on one person. So in the future we will definitely try to parallelize more tasks.
What's next for coronavirus-tracker
Some features weren't finished, mostly related to the simulation of the future spread of the virus. All of the math is done and the spread is modeled but we didn't get to translate that data effectively to the application. The next stage for the tracker would probably be to expand the API to include simulating days of Coronavirus spread in the future.