1/15/2021 5:55:31 AM
Statistics and CS sophomore Sidharth Rajaram creates COVID-19-by-county API for app developers.
Illinois Statistics and Computer Science sophomore Sidharth Rajaram is big on applying his technology expertise to public safety issues. In 2019, his hometown of Cupertino, California, recognized him with a civic award for using his ham radio operator skills as a volunteer with a disaster preparedness team.
This past summer, Rajaram began thinking about ways to apply his CS knowledge to helping people safely navigate the new normal involved with COVID-19 and mandated mitigation efforts. By fall, he landed on the idea of a website, where users could quickly find out from hour to hour how crowded or busy a particular business or location might be, so they could decide when it was safest to visit.
Rajaram developed the Plansafe backend software that takes into account a destination’s current and historical popularity and wait times by the hour. When a user types in a specific location—grocery store, hair salon, or pharmacy, for example—an algorithm calculates a PlanScore, indicating how safe the destination is at any given time.
According to Rajaram, his efforts to improve the calculation for establishments nationwide was hampered by a lack of free methods to quickly get information about all the counties in the United States.
“There was no easy way to get the data for individual counties,” he said. “It was kind of a fractured landscape for that data. All the different counties have their own dashboards or way of distributing the data.”
So, Rajaram built his own solution. The Plansafe COVID-19 county application programming interface (API) is free and is tailored for app developers. The API collects COVID-19-related data from a variety of sources, including individual states’ dashboards, Johns Hopkins University, the New York Times, and the World Health Organization.
If his sources provide conflicting data, Rajaram designed a way to arbitrate the differences. His method looks at the ratio of correct information from each source in the past and selects the source that was most often correct.
“It’s like statistical arbitration to choose which source works out in the end,” he said.
His Plansafe API is different from other COVID-19 APIs. Rather than just providing raw, cumulative data totals like deaths or confirmed cases, the Plansafe API provides a deeper level of information —such as real-time fatality ratio, incidence rates per 100,000 people, and number of active cases.
During the last two months, Rajaram has promoted the API on Reddit and at nationwide hackathons such as HackDuke 2020. Ultimately, he’d like for app developers to know this tool is available for their use, too.
“I think it would be cool if other people who want to develop beautiful front facing apps or visualizations can use my API to make the data burden easier on them,” Rajaram said. “It can also be used by individuals who are working on a research project. They can quickly retrieve data by day, and they won’t have to dig through spreadsheets like they may be doing now.”
One unique feature that developers will appreciate is that his API gets faster with each query, despite the fact that there are such a large number of days and counties for which data can be requested.
“Oftentimes what happens is that people may ask for something random like Cook County, IL, October 12, 2020,” he explained. “Anytime a new date and location is requested, that data is cached or saved, meaning all their requests around that day and area geographically will be much quicker to find [next time].”