Overview of the ALERT app
Overview of all available widgets
Uses machine learning to suggest the best times to run errands and estimates the level of risk
Shows the statistics using GEOTAB's mobility dataset
Allows for ongoing tracking of users' locations
Allows users to report having COVID-19
Alerts users on their home screen
Over the past year, governments around the world have scrambled to combat the spread of COVID-19. Although many have imposed lockdowns and attempted to put in place measures for contact tracing, the rising cases of COVID-19 prove that something isn’t quite working.
What it does
Our solution is “ALERT” - a comprehensive, data-driven solution to help keep people safe, powered by GEOTAB’s COVID mobility dataset. First, it enables users to input the location of where they’re going and sends an alert if it is a high risk area. It uses parking data to estimate the number of people active in the desired location and classifies the level of risk as low, medium or high. In addition, if the user has parked at any location where a COVID case has been reported, this data is used to alert the user. ALERT also enables ongoing tracking of users’ locations through a bluetooth connection, which addresses the problem of users not inputting their own data. Our traffic analysis on the city congestion dataset is used to suggest the best time to run the errand and is powered by our machine learning algorithm. The app cross references the store hours from Google to ensure the feasibility of the suggestion. If the user clicks on the “Statistics” widget, they can see the hourly historical city congestion data on the respective day of the week that we processed in Python to come up with the suggested time. Most importantly, ALERT enables users to report having COVID-19 to ensure that contact tracing can be carried out.
How we built it
The datasets used were provided by GEOTAB, and by querying GEOTAB's platform using SQL conventions, we gained relevant subsets of the dataset. These subsets were wrangled in Python to produce the analytics that users can see on the "Statistics" widget. Machine learning is used to the provide suggested times that are shown on the "Input Location" widget. The proof of concept walk-through of the app was created using Adobe XD.
Challenges we ran into
We ran into a few challenges wrangling the dataset, which limited the scope of the analytics we hoped to provide.
Accomplishments that we're proud of
We are proud to have provided a solution to an incredibly relevant and critical global challenge.