About the Site

Row

DataEpi: UM Epidemic Data, Analysis, and Tech Center

A resource for COVID-19 modeling built and maintained by the University of Michigan EpiMath team.

Row

What We Do

We work on modeling, data analysis, and technology to inform and assist in pandemic response.

Resources We Provide

Here you can find different resources including models, infographics, and reports about the COVID-19 pandemic. We focus mostly on Michigan but also have some resources for the Midwest region. Explore the menu above or see below for some examples of the work we do.

Row

Modeling

Our team has developed several models for understanding and predicting COVID spread:

Interactive Tools

We build apps, tools, and explorables, including:

Infographics

small image of an example infographic, illustrating the stages of COVID

We develop a wide range of infographics and brief reports to help make sense of new information about COVID-19.

Row

MI Start Map

logo for MI Safe Start, which is a computer power button with an image of Michigan inside of it

MI Safe Start Map: Track the risk levels of COVID-19 indicators

MI Symptoms

logo for MI Symptoms, which is the word MI SYMPTOMS in all-caps. The O in SYMPTOMS has been replaced with an icon of smartphone with a pulse inside of it

MI Symptoms: A free online tool built to help organizations screen their members for COVID-19 symptoms

Michigan COVID Mapping

logo for Epibayes Lab, which is an outline-mesh of lines in the shape of Michigan. A star-shape is located where Ann Arbor is.

Michigan COVID-19 Mapping: Putting COVID-19 in its spatial and social context

Recent Updates

Row

Modeling

Ridge Regression Modeling updated

  • Michigan and Midwest data updated through 03/22/2021

  • MERC-region data update through 03/07/2021

UM Modeling updated

  • CHRC slides added

State and Regional Models updated

  • Stay-at-Home model added

  • Pause-to-Save-Lives model added

Row

Interactive Tools

Interactive tools added April 2021

Ridge Regression Models

Row

About the Models

Model forecasts are generated using a ridge regression model, which is a machine learning approach that ‘learns’ patterns based on previous case, death, mobility, and vaccination data. This model has also been implemented at the national and state level as part of the CDC Forecasting Hub (see also the forecasting hub visualizer). Code and methodological details for the ridge regression model are available on GitLab.

Forecasts generated as of 07-12-2021. Case and death data are sourced from the JHU repository, and are plotted by report date.

Row

Ridge Regression Models - Michigan

7-Day Average Cases per Million

7-Day Average Deaths per Million

7-Day Average Cases

7-Day Average Deaths

Row

Ridge Regression Models - Midwestern States

For ease of viewing, confidence intervals are not displayed on the per million plots. To view the confidence ranges, please see the non-per million plots.

7-Day Average Cases per Million

7-Day Average Deaths per Million

7-Day Average Cases

7-Day Average Deaths

Row

Ridge Regression Models - Mighican Economic Council Recovery (MERC) Region

7-Day Average Cases per Million

7-Day Average Deaths per Million

7-Day Average Cases