Analysis of COVID-19 Infection Trend and its Effects on Market Behavior and Hospital Bed Availability

By | May 15, 2020

Authors:
Hadyan Fahreza (fahreza@wisc.edu)
Saranya Putrevu (sputrevu@wisc.edu)
William Wirono (wirono@wisc.edu)
Satvik Yagnamurthy (yagnamurthy@wisc.edu)

1. Introduction

The novel coronavirus (COVID-19) is a highly infectious disease caused by a newly discovered virus similar to SARS (SARS-CoV) virus outbreak in 2002. The virus was first discovered in Wuhan, China in late 2019, and the virus has spread to 210 countries and territories around the world. The symptoms of coronavirus include fever, cough, and shortness of breath. As per today, there are more than 2 million people infected with COVID-19, with more than 100,000 death tolls worldwide, making it the fourth-deadliest pandemic ever recorded in history.

This project aims to evaluate the consequence of the novel coronavirus pandemic in several notable sectors, including social, economic, and healthcare. With no vaccines and few prevention policies put in place, COVID-19 spreads rapidly from person to person. The disease has shocked a lot of people not only in their ability to spread very fast but also in the escalation of infection around the world. Moreover, due to the panic behavior from the public, the sales quantity of certain household products skyrocketed, and stock valuation of the S&P 500 and DOW plummeted, which requires further analysis to better understand current and future market conditions. Along with the stock market, many industries have been severely impacted by the virus and individuals all over the nation are left unemployed. Furthermore, as the number of cases keeps on increasing, hospital bed availability from region to region will be addressed in this report and how this can further influence the containment (or spread) of the disease in the next 3-6 months.

The result of this report could be used to predict the trend of COVID-19 infection, job and stock market behavior, and hospital bed availability in the next three to six months. In addition, COVID-19, stock market, and hospital utilization trends in the past 4 months will be visualized using Tableau.

2. Objectives and Methodology

The objectives of this report are to visualize the COVID-19 infection trend, as well as the job and stock market behavior and hospital bed availability, which are severely affected by the COVID-19 pandemic and predict the infection trend in the next 3-6 months, as well as market movement and hospital bed availability. In order to accomplish the objectives, 3 questions will be considered during the analysis process:

  1. Question 1: How will the Coronavirus be expected to grow in the next 3-6 months?
  2. Question 2: How does the Coronavirus influence the current job and stock market?
  3. Question 3: How will hospitals be impacted with increasing cases of the Coronavirus?

Tableau will be used to analyze and visualize the datasets, the conclusion will be drawn from visualizations of the datasets.

3. Analysis

3.1 Infection Trend

The dataset used to analyze the infection trend was published by John Hopkins (https://data.world/covid-19-data-resource-hub/covid-19-case-counts/workspace/file?filename=COVID-19+Cases.csv). The dataset provided a wide range of information, from a number of cases per country to the number of people recovered. The source updated the dataset regularly, so the number we used in this project may be different from the number at the time of grading.

We began our analysis on the infection trend by visualizing the number of cases per state on a map of the United States as per April 30, 2020 with Tableau, as shown in Figure 1:

Figure 1: The number of COVID-19 cases and daily change in each state. The total cases in each state can be seen by clicking at the orange circle.

To begin with, we adjusted the size of the orange circle covering each state based on the number of cases in each state, from a small circle, which indicates fewer cases, to a large circle, which indicates larger daily cases. From Figure 1, it can be seen that New York has the highest COVID-19 nationally, which has 327,959 cases, followed by New Jersey, the neighboring state of New York, with 125,880 cases. Note that this number is the number of cases as per April 30, 2020. It has caught our attention since two states with very close proximity take more than 50% of the total coronavirus cases in the United States. It might be related to the density of its population, where the density is 414 and 1,213 people per square mile for New York and New Jersey respectively. It is also interesting that states with major international airports, such as New York, Illinois, Georgia, Florida, Texas, and California, have a lot more COVID-19 cases compared to the other states except New Jersey. It can be hypothesized that the COVID-19 may be transmitted from people traveling to the United States, then developed in the states with major hubs before its widespread in the United States.

Furthermore, to analyze the spread of COVID-19, we made an interactive visual of the map of the United States and showed the daily change of the COVID-19 cases in each state as shown in Figure 2:

Figure 2: The daily change of COVID-19 infection cases in each state. The color off-green indicates that the state has lower daily change and navy blue indicates higher daily change.

Since this is an interactive map, the user can see the changes by using the tracking button on the right side of the map. From figure 1, It can be observed that from January 22, 2020, when the first COVID-19 case was recorded, to April 30, 2020, that there is a general trend in the number of COVID-19 daily cases even though the numbers heavily fluctuated throughout the days. It can be seen that the number of daily cases in each

state increased until a certain peak, and then started to decrease. From that, it can be hypothesized that the number of daily infections in each state has reached its peak and is about to decrease anytime soon. This statement is also supported by Figure 3, which plots the daily cases from January 22, 2020, to April 30, 2020, and 10 days predictions after the day the dataset updated:

Figure 3: Daily cases and prediction. Navy blue bars are the number of daily cases and light blue bars are the 10-days prediction.

From the graph, it is known that the prediction is going to decrease from its current point. Based on these observations, it is safe to hypothesize that the number of daily cases will decrease starting from April 30, 2020.

Moreover, to find out which age is the most vulnerable to COVID-19 infections and its median in each state, we plotted the median of each state on a map of the United States as shown in Figure 4:

Figure 4: Daily incidence vs median age. The color light pink shows the states with younger age median and magenta for older age median.

From the map, it can be observed that the median of the age of people infected with COVID-19 in every state varies, with the west side of the United States has a roughly younger age median, and the east side has a roughly older age median, especially around New York which is the state with the highest COVID-19 cases. It can also be seen that the older age groups are more vulnerable to COVID-19 infection since there are more states colored in magenta than lighter pink. Therefore, the government should focus the prevention measures on people in the older age group, especially in New York and the states that surround it.

Lastly, to analyze the COVID-19 infection global trend, we will analyze the logarithmic graph of COVID-19 worldwide cases and death as shown in Figure 5 and Figure 6:

Figure 5: Logarithmic curve of global number of cases.
Figure 6: Logarithmic curve of global number of deaths.

From the graphs, it can be observed that the number of worldwide cases and death is not showing signs of decreasing, yet it showed that the numbers will be vastly increased in the near future. Knowing that, the United States and every country in the world should focus on enacting proper prevention, detection, and treatment actions to stop the  COVID-19 infection. China, South Korea, and Vietnam are among the countries which successfully prevent or slow down the COVID-19 infection in their territory.Figure 7, Figure 8, and Figure 9 show the daily cases in China, South Korea, and Vietnam respectively from January 22, 2020, to April 29, 2020.

Figure 7: Daily cases and prediction in China. Navy blue bars are the number of daily cases and light blue bars are the 10-days prediction.
Figure 8: Daily cases and prediction in South Korea. Navy blue bars are the number of daily cases and light blue bars are the 10-days prediction.
Figure 9: Daily cases and prediction in Vietnam. Navy blue bars are the number of daily cases and light blue bars are the 10-days prediction.

From Figure 9, it can be seen that Vietnam has successfully prevented the infection of COVID-19, and from Figure 7 and Figure 8, it can be seen that China and South Korea have successfully suppressed the number of COVID-19 infections.COVID-19, and from Figure 7 and Figure 8, it can be seen that China and South Korea have successfully suppressed the number of COVID-19 infections.

These countries have different approaches to deal with COVID-19 infection. Vietnam had the most effective preventive policies. Just after the announcement of the COVID-19 virus, Vietnam immediately tightened up its international entry, quickly traced the travel and commute history of people diagnosed with COVID-19 and quarantined every person they had contact with, and imposed nationwide mandatory social distancing and lockdown. As we can see, Vietnam only had two new COVID-19 cases in the past three days, and Vietnam projected to only get three more COVID-19 cases in the next 10 days.

With a different condition, China and South Korea have enacted different actions and policies compared to Vietnam. Even though China and South Korea were not as effective as Vietnam in terms of preventive policies, they have enacted policies that effectively stop the infection of COVID-19. Just after the virus started to spread, both China and South Korea did massive testing, effectively quarantined the patients, and imposed nationwide mandatory social distancing and lockdown.

While it is too late for some countries, including the United States, to implement preventive measures, effective detection and treatment actions, as South Korea and China did, may stop the COVID-19 infections. However, further analyses and considerations are needed before implementing these policies in other countries.

3.2. Stock Market Trends

Moving forward, in order to determine the stock market trends, our group utilized a dataset drawn from Robinhood, a financial services company. In addition, an industry analysis studied by the University of Illinois, Urbana – Champaign Master’s in Business Administration program and developed by Experian, a consumer reporting agency.

The Robinhood dataset takes into account the top 100 most popular stocks in the New York Stock Exchange and provides information including market capacity, popularity, and analyst ratings. (https://robinhood.com/collections/100-most-popular) The source updates regularly and we captured the data into a Tableau file.

On top of this, the University of Illinois MBA program formulated a Microeconomic scenario forecasting the net amount of jobs in the top industries in the United States. Experian provided the initial framework via an estimated degree of impact in the month of March and drew the data from the Bureau of Labor and Statistics updating monthly. (http:images.go.experian.com/Web/ExperianInformationSolutionsInc/Experian_Macroeconomic_Scenario_Webinar.pdf) The slidedeck was created on April 28th, 2020 by Experian Chief Economist Mohammed Chaudhri and evaluates credit trends and implications as well, however, our group focus was on the stock and market trends.

To begin the analysis, we created a spreadsheet of all the data given from Robinhood in regards to the top 100 stocks on April 26th, 2020. The first comparison our data representation indicates is comparing the number of recorded stocks v. the percentage of change in one month for each stock. Our group began by analyzing the most popular stocks in the market because the findings would be relevant to the majority of the population that invests in the NYSE. In addition, we are incorporating a bottom-up analysis by viewing individual stocks first and then moving towards the industry analysis.

As shown in Figure 10, the percentage change ranges from -82% respective to Luckin Coffee and 116% in accordance to Groupon. There appears to be a unimodal distribution shown with the highest concentration of stock values falling under about a 20% increase. The data shown in the figure indicates companies including Groupon, Kosmos Energy, and Tesla to have the largest percent increases over the duration of the month April. As understood, April is one of the peak months for the coronavirus and these companies accelling shows that the nation is focused on efficient energy sources in relation to Kosmos Energy and Tesla. Groupon allows consumers to receive numerous discounts which is a reasonable stock to invest in during a time where disposable income is less.

On the latter end of the figure, companies including Southwest Airlines and American Airlines are facing negative improvement in their stock value. Due to many countries and states urging individuals to remain in quarantine, the trend is significant. In addition, we see minimal growth in AMC entertainment for similar reasons. According to this figure, investors should focus their spendings on companies that allow their product to grow or have high usage rates during this time. The data set compartmentalizes each stock as we can view where each company falls in the percentage change graph.

Figure 10: Percent change over 1 month for Top 100 Stocks in NYSE. The blue is placed on an alphabetic spectrum and represents the number of recorded stocks with the respective percent change.

The second data representation our group chose to include via the Robinhood dataset are the top 10 stocks in the market today. The reason we found this data important to include is because these are the stocks that have a strong weight in the S&P 500 and Dow Jones in addition to perhaps our nation’s overall GDP. We included a parameter to adjust the top 10 and view accordingly between the range of 1-10 companies. As shown in Figure 11, the stock with the highest value is Amazon at $2,384/share. The other companies that qualified the top 10 benchmark were Alphabet Class A (Google), Netflix, Apple and more.

During a time where a national pandemic is forcing our nation to rely on technology, nearly all of the top 10 companies have heavy technological prevalence in our market. The necessity for technology in order to maintain communication and obtain a steady income is becoming imperative. On top of this, we see entertainment services that rely on the internet increase their value and remain at the top.

Figure 11: Top Ten Stocks and Prices in NYSE. Circles increase in size through a direct proportion with the stock value and number displayed can be altered by a parameter.

Along with the percent change and stock value metrics, the dataset provided analyst ratings composed of reviews from both Morningstar and Robinhood analysts. These values carry reliable precedence and represent the percent of analysts that would buy this stock. Figure 12 and Figure 13 analyze the bottom and top half of the analyst ratings from the top 100 stock values.

The bottom half rating graph collects all of the stock values with below a 50% analyst rating. As presented by Figure 12, companies such as Starbucks have 42% and Ford has 28%. However, unexpectedly, Zoom has an analyst rating of 36%. It is reasonable to expect this rating to go up as the stock value has increased by 17.2% and with Universities and companies using this platform for numerous purposes in communication. Moreover, Pfizer, a global pharmaceutical and medical device company obtained a 50% rating and is growing each day due to the pressing demand for masks, coronavirus tests, and more. Finally, our group observed that financial institutions such as JP Morgan Chase and Wells Fargo have values 42% and 13% respectively and notice that these institutions are not profitable at this given state of the economy. Through this analysis, we observe that goods including coffee and more are not as recommended as other companies.

Figure 12: Bottom Half Analyst Rating. Stocks that possess analyst ratings equal to or below 50% according to Morningstar and Robinhood reports in alphabetical order.

Meanwhile, the top half rating graph obtains all of the stock values with above 50% analyst ratings and are the companies that Robinhood analysts recommend investors to place their money in. As shown in Figure 13, Amazon is leading the category with an analyst rating of 96% and this company has been vital to many families during this time. Through their rapid delivery services and access to the majority of products individuals require, Amazon has proved a great benefit to society against the coronavirus pandemic. In addition, Salesforce and Facebook have analyst ratings of 90% and 88% as companies are using Salesforce to optimize their efficiency and Facebook provides a digital communication service to any individual with the internet. Through these graphs, our group continued to notice the technology boom and critical contributions major S&P 500 companies are making to society though their stock values are incrementally decreasing. Although this is the case in the present, this industry has certainly shown that companies with more technological use tend to stay afloat in the market while disposable items and goods sink.

Figure 13: Top Half Analyst Rating. Stocks that possess analyst ratings equal to or above 50% according to Morningstar and Robinhood reports in alphabetical order.

Following the analysis on individual companies, we pursued a larger scope and analysed and practiced an industry analysis utilizing thirteen of our economy’s major departments. The data set was provided by Experian and as shown in Figure 14 and Figure 15, the Leisure and Hospitality industry lost a total of 459,000 jobs in the month of March of 2020. In addition, the healthcare industry suffered 61,200 jobs which seems significantly better than the hospitality industry, however, the impacts are severe in the United States strictly in the month of March with a net change of negative 662,600 jobs.

On the latter, industries including Information: 2000 jobs and Government: 12,000 jobs have provided the unemployed with opportunities to re-stabilize and support themselves. Connecting these findings to the stock market, we notice the company stock market trends follow a similar correlation as the net number of jobs in each industry. Information specifically pertains to many Software Engineering and Product Manager positions which appears to be increasing according to the April 28th data set.

Figure 14: Impact on Workforce divided by Industry. Change in the number of jobs possessed during the month of March per industry in ascending order according to the BLS for thirteen major industries.
Figure 15: Net Jobs in all Industries. Presents a color gradient from least to most severe number of jobs and provides the national net change in the number of jobs: -662,600 in the month of March.

Overall, our group observed that careers that can be translated virtually and are affordable are lasting, however, jobs that require being present in the on-site location and surround disposable income are losing out. The stock market follows a mirrored trend and investors should be conscious of these factors. Automation is a key factor in companies lasting or falling out and forecasts from Experian are revealing which industries may be terminated due to the virus which will be introduced. Observing which companies are profitable and which to sell, individuals will maintain positive investment actions. The coronavirus unfortunately severely impacted industries and markets, however, through tactical investment practices and entering markets with job availability, society will adapt.

3.3. Hospitals

The dataset used to analyze the effect of COVID-19 on United States hospitals was published by the Institute for Health Metrics and Evaluation (IHME), a global health research center at the University of Washington (https://covid19.healthdata.org/united-states-of-america). The dataset provided information per state including the dates of peak hospital use, excess bed demands, number of beds, ICU beds, and ventilators used at the peak.

We started our analysis on hospital capacity by visualizing the peak bed use (Figure 16), ICU bed use (Figure 17), and ventilator use (Figure 18) for United States hospitals.

Figure 16: Peak Bed Use. The yellow to blue gradient displays the US age in increasing order.
Figure 17: Peak ICU Bed Use. The yellow to blue gradient displays the US age in increasing order.
Figure 18: Peak Ventilator Use. The yellow to blue gradient displays the US age in increasing order.

To begin with, we created the different sizes of circles in increasing order from fewer to greater bed and ventilator usage. From Figures 16, 17, and 18, we can see that states that have the highest bed and ventilator usage are generally towards the east coast.

Massachusetts had the highest peak bed use at 9,857, Pennsylvania was second with 9,745, and New York was leading the peak ICU bed use with 7,667. This trend isn’t surprising because it is consistent with Figure 4, where we see that states in the east coast generally had a higher incidence of cases. With a higher incidence of cases, it is logical to assume that hospitals would have increased patients and need more ventilators.

In addition, a trend we observed throughout the Figures 16, 17, and 18 was that the median age is higher on the east coast. Corresponding to the age, the east coast is where we noticed higher peak bed and ventilator use. This trend has likely occurred because those older in age are more vulnerable to COVID-19. The correlation is especially apparent when looking at Figure 18 where we can see that states with older age groups need more ventilators, medical equipment needed most frequently for older people.

Figure 19: Excess Bed Demand
Figure 20: Excess ICU Bed Demand

We began by creating maps of excess bed and excess ICU bed demand. In Figures 19 and Figure 20, we noticed that states that have a higher number of coronavirus cases like New York, New Jersey, Louisiana, Texas, and California have a higher excess demand for bed and ICU beds. This correlation has likely occurred because hospitals are not able to accomodate beds for states with an overwhelming number of cases. It can be noticed that when comparing the excess bed demand to the excess ICU bed demand in Figure 18, there are more states who experienced a shortage of ICU beds. We hypothesize that because the coronavirus cases generally require more ICU units, there is likely a greater shortage of those in comparison to regular beds. In addition, states may have a varied capacity of hospital beds and have different levels of availability at hospitals, contributing to the increased shortage of beds in certain states over others.

This can be extremely problematic because states who have a shortage of ICU beds risk coronavirus patients overwhelming ICU units and even spilling over into other beds. This could not only cause an issue for coronavirus patients, but also for patients experiencing other medical emergencies who need space in intensive care units. States may need to continue social distancing measures to ease the burden of the virus on healthcare systems.

Figure 21: Number of States Per Date of Peak Hospital Use

We concluded from Figure 21 that many states have already hit their peak hospital use in April, meaning that many states have already started to see a decrease in the number of admitted patients. We also noticed that beginning on April 24th, there was a spike in the number of states who hit their peak hospital use. Since there is an increase of states who hit their peak hospital use in late April, we can conclude that the remaining states would follow the trend and will likely hit their peak very soon too. As a result, we hypothesize that hospital systems will soon have to handle less cases. This correlation is also consistent with Figure 3, which predicted that the number of daily cases will start to decrease. Largely attributed to social distancing measures and improvement in healthcare efficiency, we can see that many states have either already, or will soon reach peak hospital use, meaning hospital utilization will decrease.

4. Conclusion

After doing this report, we realized how fast COVID-19 spread and how much damage COVID-19 has caused in our society. In the United States alone, COVID-19 has infected 3,256,694 people and killed 233,385 lives as per April 30, 2020, with adults and elders as its majority victims. From the analysis, it is known that cities with major international airports have more COVID-19 cases than the others, which may show that COVID-19 developed in cities with international hubs. From the analysis, it is known that New York had the most infection cases in the United States with 327,959 cases, followed by New Jersey with 125,880 cases. Interestingly, these two neighboring states take up more than 50% of the total cases in the United States.

Fortunately, based on our 10-days prediction, the COVID-19 infection cycle is already on its peak and is expected to decrease in the near future. Knowing that, strict prevention, detection, and treatment actions are needed to ensure that the infection cases will decrease. There are some countries that have successfully pushed down the infection cases in a short time with their own policies. These countries are China, South Korea, and Vietnam. The United States government should reflect on these countries to effectively prevent the COVID-19 from spreading even further.

Furthermore, from our analysis, we found out that the COVID-19 pandemic significantly affects the market behavior, especially stock market. As shown in the data, many companies are experiencing minimal growth and recovering from a significant downfall from the month of March. The companies that are in the top 10 are all specialized in consumer marketing or digital technology including Amazon, Google, and Netflix. Rather, various other industries are facing the risk of becoming extinct as the market portrays them as relatively disposable.

On top of the stock market, the coronavirus has heavily affected the job market as industries including leisure and hospitality in addition to retail and trade have experienced high unemployment rates. However, information and the government industries have been able to provide jobs to the people in the month of March. Through our research and data analysis, we found that companies that have essential services are surviving and in this economy, technology is no longer a virtue but a fundamental necessity. In the future, we predict that automation, computer and data science, and virtual financial services will outlast retail industries. The coronavirus has certainly provided a unique scope of the fields and jobs that will be extinct along with careers that will be born post COVID-19.

Lastly, the COVID-19 has greatly impacted the healthcare system. Especially in states where COVID-19 cases are especially high, we found that there is an excess demand for hospital

beds that not all states have been able to provide to patients. However, since many states have hit their peak for hospital bed usage in late April, we predict a brighter future where hospitals will soon see a decrease in patients. We acknowledged that hospital bed availability is vital to treat patients with COVID-19 and to ensure effective patient care, states will need to continue social distancing measures to ease the burden on the healthcare system, while hospital systems continue to construct a game plan to handle the virus.

Share this post
  • 9
  •  
  •  
  •  

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.