Output from the program SAS; has lines of code outlining how we pull data from a statewide database
Example of SAS code used by the Data Team

During the course of this response, we have created hundreds of fact sheets, reports, and guidance documents, but our most popular resources continue to be those created by our Data Team. Each week they create the massively popular Data Snapshot. Every day they update the data dashboard. But what does the rest of their day look like? We wanted to share a behind-the-scenes look at our Data Team and the work they do. Here’s what happens on a typical weekday:

7:45 AM – Regroup

The team regroups from the night before. Often, they’ll have dozens of emails waiting for them—data questions, other information pertinent to the overall response, and updates from CDC and the Wisconsin Department of Health Services (the state health department).

8:00 AM – Meetings

One representative from the Data Team attends the Operations Team meeting, where they learn about themes emerging from contact tracing which prompts the Data Team to pull data to assess trends. For example, hearing many cases describe going to particular bars can lead to the team developing code to capture cases that referenced going to a bar in their interview.

The Data Team Lead attends the Incident Command Chiefs Meeting, where updates across the response are shared and overall strategy is established. Data-informed decisions, such as testing strategies and messaging, are made on a daily basis.

8:45 AM – Prep the Dashboard Update

Every day, including weekends and holidays, a member of the Data Team updates the data dashboard. Over time, this has gone from just a handful of data points to 25 different data panes (see evolution of dashboards at bottom of page). They regularly adjust based on feedback we receive from the public and work with partners at the City of Madison IT Department as well as ESRI to make the changes.

They download case information that our contact tracers and case investigators input into a state database called WEDSS (Wisconsin Electronic Disease Surveillance System). WEDSS has been in existence for a long time. It’s used to track every reportable communicable disease from Lyme disease to last summer’s vaping-related outbreak to COVID-19.

The Data Team then uses statistical programs called SAS and R to translate the raw data into usable data summaries. In order for these programs to work, our Data Team has to write and run many lines of code to calculate and summarize all the data we need. To run the daily update for the dashboard takes around 1,000 lines of code in SAS (see image at top of page) and 650 lines of code in R. Once the data summaries are exported into Excel, the team member hand enters each number into the dashboard interface. This is a tedious task! It takes about 30 minutes to update all the fields in the dashboard.

10:00 AM – Daily Huddle

Every weekday the team does a daily huddle to talk about methodologies for calculating new data, the latest trends, and other considerations for the day. Sometimes these meetings have guest appearances from a member of the Operations Team (these are the folks that do everything from interpreting state and federal guidance to contact tracing), the Liaison Team (these are the folks that serve as a link between the health department and sectors like schools, parks, businesses, and community centers), or the Communications Team. We like to cross-pollinate teams so everyone is in the loop and can problem solve together.

11:00 AM – Provide Consultation to Communications Team

Folks are very interested in the data, and it means we receive a number of questions every day on social media and from reporters. We get questions about how to interpret certain data points or why we use one measure instead of another. When a Communications Team member gets a question they can’t answer, it often means a quick Skype chat or email to a Data Team member to hammer out a response.

11:30 AM – Prep the Data Snapshot

The weekly data snapshot is released on Mondays (except for days when we announce new orders). We wait until Monday to run the data for the snapshot because it includes data through Friday. Lately, the snapshots have focused on updating metrics that guide our progression to new phases of reopening. Like with the daily dashboard update, this requires a substantial amount of code (~1,300 lines), and there is lots of double-checking and confirming among team members to make sure everything is correct. Once the metrics are calculated, they spend time with the Communications Team perfecting data visualizations and framing language in order to help people understand and contextualize the data.

1:30PM – Analysis!

We analyze data in a very collaborative way—never in isolation. A Data Team member will suggest an approach, and it is then modified, tested, and validated by the other team members. This often includes doing a quick review of the most current scientific literature—how have state, federal, and international analysts looked at this question before? Which variables are other local health departments or the state health department in Wisconsin using?

Some examples of recent questions we’ve received: What is the median number of contacts of symptomatic cases compared with pre/asymptomatic cases? How many cases have there been in each Dane County municipality? How many cumulative cases have there been between the 0-5 age group? Can you map census tract level testing rates superimposed on neighborhoods with lower rates of health insurance?  When they investigate the answers to these questions, they often double-check—is this a ‘nice to know’ question, or will this inform action? They prioritize analyses that inform action.

2:00 PM – Weekly Department of Health Services COVID-19 Webinar, or Consultation with Other Public Health Staff or Researchers

The state health department holds a weekly COVID-19 webinar for local public health staff where they provide updates on various topics, such as WEDSS changes, methodology changes for calculations, or guidance updates on case investigation and contact tracing. Data Team members, along with many other Public Health Madison & Dane County staff, regularly listen in to this webinar. This is also an opportunity for staff to ask questions of the epidemiologists from Wisconsin Department of Health Services.

Data Team members also regularly communicate and work with analysts at the state health department, other local health departments, and UW researchers. They are constantly providing feedback to one another to make our data more understandable, consistent, and transparent.

3:30 PM – End of Day Briefing

Multiple times per week, all 100 or so staff members involved in the COVID-19 response has the opportunity to attend a virtual end of day briefing. This is where section chiefs share the latest news and answer questions of staff. The Data Team lead shares an update here as well and loves to share their screen to illustrate the latest trends we’re seeing.

5:00 PM – Done for the Day (…Hopefully?)

Even though the day is done, they’re never really “off call.” For example, if our health officer needs some information urgently to inform a decision, they jump to action. COVID doesn’t run on regular business hours, and neither do we!

We are extraordinarily proud of the Data Team for their talents and reputation as leaders in epidemiology throughout the state. Because many counties do not have the capacity to staff a data team, our team often lends their skills and expertise to others. The team is also able to provide timely feedback and local perspective to the state health department as they consider changes to their systems as well.

The Data Team is small but mighty, with three epidemiologists/analysts and one team lead. As the team puts it: “Our days are busy, but we’re truly grateful we get to work on the response. There are few thing within anyone’s control during this epidemic. To live in a county that recognizes the value of utilizing the data we do know is a privilege.”

Iterations of the Data Dashboard

Image of the first iteration of the dashboard
First iteration of the data dashboard
Image of the second iteration of the dashboard
Second iteration of dashboard
Image of the third iteration of the dashboard
Third iteration of dashboard
Image of the current iteration of the dashboard
Current iteration of dashboard
This content is free for use with credit to the City of Madison - Public Health Madison & Dane County and a link back to the original post.