Inner County Heatmap
COVID 19 Detection – Team 05
William Thesken, Olivia Lazaro, Khoi La, Nour Kahile, Gregg Bushyeager
July 24th, 2020
COVID-19 created a need for localized testing data. Indiana is fortunate in that state-wide data is easily accessed. Companies like Regenstrief and government agencies like The Indiana State Department of Health can then use this data to display the state’s COVID-19 status. Heatmaps are among the many visualizations used to convey data and are a way to display concentrations of positive test cases. Current efforts in Indiana display positive case counts at the zip code level. Moreover, other states are not as interoperable with their healthcare data. The proposed solution aims to remedy the challenges just described.
There are many proposed solutions already on the market, like the Google/Apple effort (Apple, 2020), but these have controversies associated with them. Namely, many Americans are concerned with data privacy which is in direct conflict with the goals Google/Apple set out to achieve. Other efforts require individuals to self-report; there are certain ethical concerns with this solution. The team’s proposed solution will circumnavigate these concerns.
The solution maps occurrences of positive tests to Marion county test centers. This will provide users with a map of high-risk areas at a fine granularity, which is particularly useful in densely populated areas. An example accompanies this document. Two major assumptions are necessary for this solution to work. The first assumption is that the reported data includes test outcomes faceted by test location. The final assumption is that people will seek the most convenient testing center when getting tested, whether it be near their home or work. The importance of the second assumption is that for the data to provide useful information, the test outcomes should be associated with the testing location’s area.
Initially, the team only identified one customer segment. COVID-19 is known to affect those with preexisting medical conditions and of old age (Centers for Disease Control, 2020), who will be referred to as a vulnerable population throughout this paper. Based on this notion, the team concluded this subset of the population may have an interest in the product. After conducting multiple surveys, younger individuals were identified as a second customer segment. It is speculated that these populations would be key customers as they are more likely to need an inner county heatmap for health purposes.
The survey was primarily used to identify and confirm customer segments. Results from the survey are summarized in the graphs that accompany this document. The main and only relevant takeaway from the data was the identification of a second customer segment. The surveys were unable to confirm early speculations about vulnerable populations.
Ideation and Iteration
The team contemplated several different solutions before ultimately deciding on the heatmap. Within the team, originality was paramount; this served as a dissuasion from the cliché contact tracing apps that already existed. The first proposed solution was aimed to aid the homeless and volunteers. This solution would provide a medium for volunteers to help where it is needed most. After moving on from that topic, a grocery delivery application was considered. This too was turned down; both solutions did not align with the expected project criteria.
After pondering for several days, the team ultimately fell back to a contact tracing or outbreak detection map, though originality was still heavily valued. The decision was then made to move forward with the heatmap at an inner county level, as depicted in the prototype. There were many iterations and changes made to the project as the team gathered insight. The prototype underwent numerous design and functionality changes; the goals proved to be too lofty. The second prototype needed tweaks to the descriptions in addition to other design changes.
PRO Squad experienced several challenges along the way. The team had to learn how to connect the server with the database. API keys were finicky which postponed development. Embedding and interacting with the map was challenging; ultimately the team could only embed the map. PRO Squad learned about several different topics ranging from RESTful APIs to AGILE project management.
Project Management and Product Development Tools
Various tools were used for communication and coordination purposes throughout the project. The team used slack to communicate with one another throughout the project’s duration. Initially, Trello was used to document research and share ideas. The team eventually required a more comprehensive project management tool and switch to the Confluence platform. Confluence not only offers a centralized hub for the team, but it also features Jira and Bitbucket, two useful tools for product development.
The GO squad used a few notable tools to facilitate the ideation process. Miro was used to brainstorm synchronously. The data visualizations were created with Tableau. Go Squad also used Draw.io to aid in the visualizations. Pro squad used Jira to manage the development of the product. The web application was hosted with Azure services. The team used C# and Asp.net Core and wrote in Microsoft Visual Studio Pro. SQL was used to query and store mapping data, counts, and locations. Unfortunately, we couldn't obtain the necessary resources to export our map widget. The self-contained server we used doesn't support complex applications of this nature. PRO Squad used Microsoft SQL Server Management Studio to populate the table to store our data and IDs. The team used R to display test data on the county map. The dropdowns were made with CSS.
Development Process and Description
A short chronological description of the project's process is detailed below.
- Brainstormed relevant issues in the community related to COVID-19 and the associated proposal
- Lack of inner county outbreak maps
- Inner county heatmap
- Need for sick individuals to stay home
- Grocery delivery service
- Need medium for volunteers to help those most affected by COVID
- Homeless helper
- Lack of inner county outbreak maps
- Decided on the heatmap
- Other options were off-topic
- Developed three iterations of a potential product and reached a conclusion
- GO Squad identified the customer segment and every other deliverable by the end of week 3.
- PRO Squad began work on the project at the beginning of week four.
The two squads operated differently. The GO squad worked synchronously on the same part throughout the entirety of the process. PRO squad partitioned their work into different areas of the product.
Below is the product's functionality from a technical standpoint.
- Dr. opens the website and registers the testing location
- Azure handles the username and password data
- Azure handles the location data
- Dr. reports the number of total and positive cases for the day
- SQL database stores this information
- User opens the website to check their locality
- R makes SQL calls to the SQL database to display the testing information on the map
- R makes SQL calls to the Azure database to retrieve
To identify the customer segment and the viability of the product, the team conducted three phases of surveys. The first phase simply asked 20-30 people to critique and review the team’s prototype. The feedback received at this stage was invaluable. A team member dispensed a quick two-question survey to assess the viability and practicality of the product; it was received well. A final survey was utilized to further segment the population for marketing purposes. This survey was the largest with 60 respondents; some rudimentary metrics are shown in the presentation accompanying this document.
A final metric used to identify the NPC is well documented on the business model canvas.
If given another 5 weeks to work, it would be a goal to be able to expand the scope of the data to multiple counties in the surrounding area specifically. Because the setup of the data analysis is primarily concerned with emulating hotspots through a large number of testing sites in Marion specifically, it may be less effective overall to apply the same principles to less densely populated counties, but with a data set taken as a representative sample from the widely-distributed testing centers in Marion, having a total of 28 testing centers, this would likely be doable for other urban centers in other counties both in Indiana and other states. Another goal we originally had was a 'live update' running on constant data refreshing from the contributed data to the database, something that would likely be implementable if true data was contributed daily, but would require additional work to code in the connection. Overall, what we wanted to accomplish as a basic requirement was met, and any expansions in the future for implementation have been planned for implementation, and any changes that would be made with more time would likely just be polished.
Nour Kahile (PRO Squad)
Nour is a front-end development team member that was responsible for implementing the drop-down boxes for the county, location, and provider. He designed the login and registration system for doctors as well. Nour wrote part of the 'Product Development Tools' section in this document.
William Thesken (PRO Squad)
Olivia Lazaro (PRO Squad)
She was responsible for R coding and ideation for the finalized product. Olivia did research on the available information for the COVID testing centers and R software and coded the map embed as well as plotting the points on the map itself. Olivia wrote the 'Looking Forward' section in this document.
Khoi La (GO Squad)
Khoi conducted market research. Largely responsible for the completion of all the deliverables:
- Empathy map
- Business model canvas
- Customer personas
- Value proposition
- Environmental analysis
Gregg Bushyeager (GO Squad/ Project Manager)
He conducted most of the market and product research; he was also partially responsible for the deliverables. Aside from setting up the project management tools and coordinating the team, Gregg prototyped the product three times, created and distributed detailed surveys, performed rudimentary data analysis, obtained site coordinates and addresses, helped with the final presentation, and wrote most of this document.
Apple. (2020). Privacy-Preserving Contact Tracing - Apple and Google. Retrieved from https://www.apple.com/covid19/contacttracing
Centers for Disease Control. (2020, June 25). People Who Are at Increased Risk for Severe Illness. Retrieved July 23, 2020, from https://www.cdc.gov/coronavirus/2019-ncov/need-extra-precautions/people-at-increased-risk.html