In these uncertain times, the W. Capra Data Science team has taken upon itself to report on the anticipated impact that the COVID-19 outbreak will have on the verticals we service over time. Using data current as of April 13, this first report is intended to identify and predict how the virus will develop by state, which states will successfully contain the outbreak fastest, and which other potential factors or time scales might be notable in modeling the impact of the outbreak. The following is a summary of our findings. The full report can be found here.
A note before we begin: This outbreak and the data surrounding it changes daily. This report was created when looking at the outbreak as a data problem that might benefit from data driven solutions and insights. It is not intended to be a substitute for medical or safety advice, nor is it a recommendation on outbreak response currently in place in various locations around the country. Individual assessment of local laws and current official government and health guidance should be reviewed before making any decisions.
The current data for the United States included information to construct a first look at how states were handling the novel coronavirus outbreak and build an initial view of important trends across all states. In particular, the important trends examined were as follows:
- Testing Volume and Capacity
- Case Rate
- Death Rate
- Case Velocity
- Death Velocity
- Case Acceleration
- Death Acceleration
Many other factors such as population density, social distancing success, healthcare networks, and others also affected each state-by-state comparison. In order to evaluate states against each other, the various inputs will be equalized on a per-capita approach.
Modeling the Stages of COVID-19
For the outbreak, there are four main stages named as following: Exponential, Linear, Improvement, and Containment.
- The Exponential stage sees the number of new daily cases and deaths increasing day over day, a positive velocity and acceleration for cases and deaths. This stage is where the hardest-hit locations around the country are now.
- The Exponential stage transitions to the Linear stage, where the number of new daily cases and deaths are flat which indicates still a positive velocity but an acceleration approaching zero. For now, the goal is to see states reach a point where the change in daily new cases and deaths is zero or begins to decrease – a transition from stage 1 to stage 2.
- The Linear stage then becomes the Improvement stage where the number of new daily cases and deaths is decreasing – indicating a decreasing velocity with negative acceleration.
- Eventually, these will all give way to the Containment stage where there are zero or near zero number of daily new cases or deaths. The various trends listed above will help to determine when these velocity and acceleration changes occur and aid in the prediction of who might reach the different stages next.
Testing the population is the first step to containing the outbreak in each state across the United States. However, testing numbers have been small and unevenly distributed across the US. Only 0.9% of the entire US population has been tested, with some states – like New York, Washington, and Louisiana – testing a much larger proportion of their populations (over 2,000 per 100,000) compared to others – like Texas and Georgia – where testing is below 500 per 100,000.
Monitoring the case and death rates and rates of growth moving forward will show which states have successfully implemented various strategies to combat the outbreak and which states are trending in either positive or negative directions.
Case rates, or the rate at which COVID-19 tests return a positive result, vary greatly across the country. High case rate states like New York and New Jersey can be attributed to higher population densities and the ease of transmission of the infection; other high case rate states like Michigan and Oklahoma, despite their low population densities, points to lower testing rates geared toward symptomatic patients in these locations.
Similar trends occur in the death rates, or the rate of positive cases that result in death, of states. While the US overall has a death rate of 2.23%, states like Washington and Michigan have much higher death rates compared to much of the rest of the country as the disease has hit different populations such as those of advanced age or the prevalence of health conditions among residents.
The growth rates of both cases and deaths is still in its early stages. Currently, the positive case growth is directly tied to testing numbers and availability. However, transmission rates in states can lead to understanding important factors for the spread of the virus.
In order to normalize time across the different states, the initial date is called the “infection day” or the day in which the state first has more than 100 positive cases. For example, the data says that California has not experienced as much growth in cases as New York since infection day, but a large underlying reason for that may be that California is testing at 1/5 the rate of New York per capita. Additionally, Washington is testing at almost half the rate of New York but has a much lower case rate and rate of growth than New York, which points to the success in Washington of various measures such as a statewide lockdown and social distancing.
Similar results in the change of the death rate have been measured. Deaths in California and Washington and growing at a slower rate which can be also be attributed to lockdown and social distancing efforts. All the indicators in states like WA and NY point to their Case Velocity peaking soon.
All states are currently in either the Exponential or Linear growth stages; however, states like NY and NJ have seen a recent negative trend in their accelerations, which indicates that they and others may have peaked in terms of cases. Other regions – such as the Pacific coast and New England – appear to be following similar case trajectories. Specifically, California and Washington have had success in flattening their curves with the slower growth rates per capita. These successes point to the significant success in applying and adhering to lockdowns and social distancing measures.
Assessing the Business Impact
These findings are powerful for their insights into various industries, especially those that are consumer focused.
Retail, an industry worth over $3.5 trillion dollars in 2018, stands to undergo major changes from the fallout due to the outbreak. This crisis is unlike anything seen before by the industry. According to Retail Dive, total retail traffic fell 9.1% from the beginning of March, with that number being buoyed by the increase in retail traffic at stores that sell essentials. The estimate for non-discretionary retail stores is nearer a 30% decrease in traffic. Looking at similar nationwide events in recent memory details when the retail market may look to rebound. Retail sales restarted after six weeks following 9/11 and retail sales took a full six months to restart following the 2008 recession. Due to the unknown end date of this outbreak and still unforeseen economic consequences from the outbreak, more data is imperative to know when retail can restart and how its customers and employees might resume normal life.
Despite the heavy losses to retail, the E-commerce space stands to break out albeit under precarious circumstances. Although E-commerce will not fully recoup the losses sustained, new opportunities have been presented with changes in how the customer base behaves. 52% of consumers are avoiding crowds and 32% of consumers are leaving their homes less often. These drops have resulted in customers reaching out for new ways to get the things they want and need. 58% of surveyed consumers have purchased more items online compared to in-store than they normally would have. The results from this report and reports to follow will look to see when retail can restart and will help to monitor how the e-commerce space has changed as the consumer base has adapted in the outbreak.
The petroleum and energy industries have a large stake in the outcomes of this outbreak. The problems in the petroleum sphere are two-fold: travel restrictions and the slowing of supply chains. In the short term, the personal travel limitations and hurt revenues with jet fuel and reduced the everyday consumers use of oil and oil-based products. Looking at China and its response to the outbreak, including how it begins to lift its travel restrictions and lockdowns, will give insights into how much longer the restriction of travel in this country will last and what the industry can look to with regards to the timeline for recovery. In the hardest hit areas of China, the full lockdown is beginning to be lifted nearly 3 full months after implementation with industry and travel following closely behind. This is after many weeks of these locations reducing their number of overall cases and numbers of daily new cases which is very far away for many regions across the US.
An Ongoing Assessment
The W. Capra Data Science team is continuing to develop its framework for evaluating the severity of the disease and the success of states around the country in combatting the disease. Various factors and trends will be examined moving forward. The trends listed in the initial report will continue to be the benchmark performance metrics for the states over time and will reflect positive and negative trends across the country. In addition to the trends already stated, other developments are in the works. Refining the scope of the outbreak to a county/parish level from a statewide view, including weather and mobility data in the trends, and predicting future velocities and accelerations are all important additions for upcoming reports on COVID-19 in the United States from the W. Capra Data Science team.
For further discussion of data modeling or anticipated COVID-19 business impacts, contact the W. Capra Data Science team:
Nate at firstname.lastname@example.org
Stu at email@example.com
Data from The New York Times, based on reports from state and local health agencies. (2020, April 6). Retrieved April 3, 2020, from https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html
The COVID Tracking Project. (2020, April 6). Retrieved April 3, 2020, from https://covidtracking.com/
Coronavirus and the oil market: The effects thus far and what to expect next. (2020, March 19). Retrieved April 3, 2020, from https://www.atlanticcouncil.org/blogs/new-atlanticist/coronavirus-and-the-oil-market-the-effects-thus-far-and-what-to-expect-next/
Hjelmgaard, K., Lyman, E. J., & Shesgreen, D. (2020, April 1). This is what China did to beat coronavirus. Experts say America couldn’t handle it. Retrieved April 3, 2020, from https://www.usatoday.com/story/news/world/2020/04/01/coronavirus-covid-19-china-radical-measures-lockdowns-mass-quarantines/2938374001/
Davis, D., Demery, P., Freedman, L., Crets, S., & Brohan, M. (2020, March 19). Coronavirus will boost ecommerce in the long run, but brings new risks. Retrieved April 3, 2020, from https://www.digitalcommerce360.com/2020/03/19/coronavirus-will-boost-ecommerce-in-the-long-run-but-brings-new-risks/
Coronavirus effect: Bankruptcies won’t save retail this time. (2020, April 5). Retrieved April 6, 2020, from https://www.cnbc.com/2020/04/04/coronavirus-retail-is-used-to-existential-crises-but-bankruptcies-wont-save-it-now.html
Jansen, C. (2020, March 18). Retailers brace for the impact of coronavirus. Retrieved April 3, 2020, from https://www.retaildive.com/news/retailers-brace-for-the-impact-of-coronavirus/574346/
Petro, G. (2020, March 20). The Coronavirus Tsunami: What’s To Come For U.S. Retail. Retrieved April 3, 2020, from https://www.forbes.com/sites/gregpetro/2020/03/20/the-coronavirus-tsunami-whats-to-come-for-us-retail/#61f1f7e5609d
Thau, B. (2020, March 20). Coronavirus Effects on Businesses in Major Industries. Retrieved April 3, 2020, from https://www.uschamber.com/co/good-company/launch-pad/coronavirus-effects-on-major-industries