Github: marcusze/covid-tracking
Background
More than 1/3 of the retirement homes in Sweden report Covid-19, despite having strict visit restrictions and health controls of employees.
The objective of this project is to identify asymptomatic Covid-19 carriers among employees in retirement homes quicker and more accurate.
Solution
We find inspration from fraud detection of skimmed credit cards.
The analogies are
Credit card fraud | Covid tracking |
---|---|
Credit card transactions | Patient visits |
Fraudulent transactions | Infected patients |
Skimmed terminals | Asymptomatic carriers |
Inputs:
- Visit logs
- Infected risk patients / elderly
Output(s):
- Ranking of most probable asymptomatic carriers among caretakers
- Probability of future infected patients
Screenshots
Check out the detailed project decription, illustrations and screenshots in the shared pdf here.
Illustration of the search algorithm
Data model from simulations
Next steps & Contributions
- Live trials and feedback: Looking for collaboration partners to conduct real searches on real data. They will be able to customise the solution to their needs.
- Algorithm and data model improvements: Open for suggestions
- Tech improvements: Open for suggestions