The risk group (notably the elderly people) is exposed to a higher probability to suffer of severe complications if they contract the virus. Our goal is to advise them safer routes in order to decrease the risk of being sick.

What it does

Since the data about the persons are generally subject to ethical critics, our goal is to simulate the movement of the population under different scenarios, using as much acceptable information as possible (e.g. the population density, the number of residents, etc.), in order to define, as precise as possible, the less frequented areas where the risk group does not encounter a major risk of contamination. This is a probabilistic approach and can be enhanced, for instance, by training the model with real data and then only expose the predictive/estimated (and not live) scenarios. This has the advantage that the real data would not be used for real-time tracking or be exposed by any sort to ensure the privacy of everybody. The software then recommends a safer route for the risk group.

How we built it

Concretely, the program simulates a city in activity with agents representing the population (proportional to the numbers of inhabitants) under different scenarios (e.g. quarantine, immunized ratio, stay-at-home rate, etc.) and shows a website with safer (less frequented) areas. The impact is to inform people at risk (e.g. the elderly people, the individuals with underlying medical conditions, etc.) where they can wander without too much risk. It is currently written in Python with Plotly and Mapbox and can easily be brought to the cloud.

Challenges we ran into

The challenges were multiple. First, the choice of the data; what pertinent information has to be taken into consideration to simulate the movements of the population. Second, the elaboration of the model; how to train a simulation generator that can reflect the reality as close as possible (and eventually trained with real data). Finally, what kind of setup can deal with a large amount of data for the simulation using cutting edge technologies (like GPUs, etc.) and can bring it to the Web.

Accomplishments that we're proud of

We were proud to accomplish in a short time everything we wanted to create; so a Web prototype, ready to be enhanced with real data.

What we learned

The development of a map with Plotly and mapbox and also how to create a video with Adobe Rush.

What's next for Safe Routes

  • Bring it to the next level with a more complex model taking into account more information and eventually trained with real data.
  • Scale it to the cloud.
  • Advertise the solution.

Try It out



dash, mapbox, python

Devpost Software Identifier