Why it is Important: We can track the patterns of current COVID-19 cases and apply those correlations into an app that’s easily accessible to the standard consumer to prevent future hotspots of disease spread. In the case of our L'Destination app, data patterns revealing higher-risk vacation spots alert users to alternative, safer locations.
How It Works: By taking the various data types, such as temperature, humidity, population density, cases counts, and map data, we can process them into a Machine Learning program that would help us find any correlations to those above factors and the risk of infection. These findings would help us calculate risk ratings for each region and would help users plan their future trips. While the risk of infection from the coronavirus still present, we can prevent users from going to higher risk areas unnecessarily.
We are team L'JAK^2S, the Jack(s) of All Trades! We are scholars from the NASA L'Space Mission Concept Academy with members across the US and Puerto Rico. We thought that the Human Factors challenge was a great way to showcase our expertise in conceptualizing future products or services while remaining creative and having fun throughout the process due to the immense amount of variables at play.
We used the provided NASA SEDAC Data and the data from the Sentinel-Hub EO Browser for different datasets such environmental factors, number of confirmed cases, population density, moisture index, and more!
We utilized various tools such as Tableau and Microsoft Excel for data analysis, Google Suite Products for documentation and organization, Autodesk Sketchbook for concept drawings, and Zoom and Discord for communication purposes.
Our most difficult problem was the question of how we could help the world's population since machine learning was not one of our team's highest strengths. Additionally, data analysis was heavily time-consuming and 48 hours was not enough to design an accurate model or a concrete blueprint of a model for the pandemic. Halfway through day 1, we were confused and uncertain of our next steps. We had started looking at how population data correlated to beaches and COVID-19 spread, but we did not have the computing skills to analyze the data in the way we wanted.
We decided to pivot. We realized that other organizations such as government agencies, private corporations, academia, and NCOs could probably create very accurate machine learning algorithms to model the pandemic, but would have issues applying their findings to the masses. Given that beaches are reopening, we wanted to find a way to use the risk analysis we had started to develop and tell people which beaches were safer than others. We designed a mock-up for an app that would do this. Our proudest accomplishment is creating a solution that uses our team's strengths in creative solutions and design that could help bridge the gap between the cutting-edge research and mass application. Our project allows those who wish to travel to at least do so safely, potentially saving thousands of lives in the process.
Our website: https://ldestination.weebly.com/
SEDAC: https://sedac.ciesin.columbia.edu/mapping/popest/covid-19/
We used SEDAC data to understand population density and cases per 100K.
EO: https://apps.sentinel-hub.com/eo-browser
We looked at EO data such as moisture and temperature data.
We also looked at local county websites for each of the beaches we analyzed.
https://corona-virus.la/SaferLA
http://www.longbeach.gov/health/diseases-and-condition/information-on/coronavirus/covid-19-orders/
https://occovid19.ochealthinfo.com/coronavirus-in-oc
We created an Excel spreadsheet with SEDAC, EO, and safety data from 45 different counties. We gave each beach a safety rating given the strictness of their social distancing requirements. We started by analyzing this data with MATLAB and Excel. We were especially curious about a correlation between cases per 100K and safety requirements and created some graphs to analyze this using some open source code (linked below.)
Excel Code: https://support.microsoft.com/en-us/help/213750/how-to-use-a-macro-to-add-labels-to-data-points-in-an-xy-scatter-chart
After going through this process, we decided to design an app that would enable users to see a risk total (based off a risk matrix that a theoretical machine learning algorithm would create). The user could then make an informed decision about where to vacation based off this risk value. While we were not able to actually create this ML, we came up with a general idea of how it would work and what it would do (shown on our website). We then focused more on designing our app.
Unsplash (Image Sources) : https://www.unsplash.com/
We used Unsplash images of beaches to help design our app and make it user-friendly and inviting!
With unlimited time and resources, we envision that a fully-developed version of the L'Destination app could enable people to visit beaches and other destinations, or even grocery stores, safely.