Stephanie Kopet's Portfolio Project
Throughout the Data Analytics Program, students are asked to imagine themselves as analysts for various imaginary companies. Below is Stephanie’s descriptive analysis of staffing during a US flu pandemic.
Hi Stephanie! What inspired you to make a career change into data analytics?
I really enjoyed the free, email short course from CareerFoundry, so I thought I’d give it a go!
What did you enjoy most about the program?
I loved developing a relationship with my mentors and tutors, and understanding how to work with real-world data.
Talk us through this portfolio project.
I enjoyed shaping the data and applying it to other areas. I also loved the research behind the data and data trends, and turning it into my own presentation.
- Project goals
- Defining the research questions and hypothesis
- Sourcing and preparing the data
- Statistical and visual analysis
1. Project goals
The United States has an influenza season where more people than usual suffer from the flu. Some people, particularly those in vulnerable populations, develop serious complications and end up in the hospital. Hospitals and clinics need additional staff to adequately treat these extra patients. The medical staffing agency provides this temporary staff.
Determine when to send staff, and how many, to each state.
The agency covers all hospitals in each of the 50 states of the United States, and the project will plan for the upcoming influenza season.
2. Defining the research questions and hypothesis
Clarifying questions and funneling questions:
When is flu season?
- Does this vary by state?
- Is the duration the same in each state?
- Will current staffing contract durations cover the entire length of time for each area’s need in staffing (typical travel nurse is 3 months)
What populations have higher death rates?
- Does this vary by state or is it consistent with vulnerable population assumption?
- Are there other factors affecting death rates in high percentage areas such as air pollution, available technology and resources, etc that affect death rate?
- What is percentage of deaths that received flu vaccine?
Which states have higher incidence of the flu?
- What are general population demographics contracting disease?
- Is one population at greater risk by state?
Do organizations that require flu vaccine have a decreased demand for additional staffing?
- Does this decrease flu contracted in hospital by patients?is staff infecting patients?
What is meant by minimal under article 2 in success factors section?
- How is this measured?
Do staffing requirements vary by state?
- Are death rates affected by this?
- Does this number account for acuity of different units?
- Do additional resources and staff- ratio percentages negate hospital pool staffing or floating staff between units? On call pool or calling off staff (how do you measure this?)
Do hospitals that employ nursing acuities have lower incidence of re-hospitalizations and better outcomes?
- Are lower costs associated with this?
- How would the staffing agency accommodate hospitals that utilize nursing acuities for staffing needs?
If we provide staffing to decrease patient nurse ratios as follows:
- August- October staffing to assist with flu clinics in areas with larger vulnerable populations, to provide vaccines and aide in handwashing, nutrition campaigns, teach good habits such as coughing into arm prior to peak season;
- September-March aide in staffing units receiving flu patients mainly med surge and critical care nurses in areas with greatest vulnerable populations and highest rates of flu visits and flu deaths from previous 5 seasons; Then, we can lower death rate to .01 and reported cases will decrease by 20% from the 2019-2020 flu season.
3. Sourcing and preparing the data
- Influenza deaths by geography, time, age, and gender Source: CDC
- Population data by geography Source: US Census Bureau
In Excel, the data was
4. Statistical and visual analysis
From the data obtained, the math conducted in the analysis, reveals there is a strong correlation between vulnerable population by state and flu deaths reported as well as the weekly report of vulnerable population flu visits and flu deaths. The two comparisons of three variables both resulted in a 1 for correlation signifying a strong correlation between the variables. The following will break down mean and standard deviation for the three variables:
Weekly flu visits from vulnerable populations: Mean – 8829
Standard Deviation- 4765
Standard Deviation- 121
Vulnerable population by state:
Mean- 1,378,981Standard Deviation- 1,501,476
There is no significant difference between flu visits and flu related deaths
H0: population mean of vulnerable population flu visits = vulnerable population of flu deaths
There is significant difference between vulnerable population visits and flu related deaths Ha: mean of vulnerable population flu visits <> mean of flu related deaths
Using a standard level of significance of 0.05 to conduct statistical analysis and finding a p-value of 5.21488 E-18, we will reject the hypothesis that there is no significant difference between flu visits and flu related deaths. This means there is a significant difference and should not be used as main factor in providing staffing even with the strong correlation between the numbers. After further inspection, it is easy to see that weeks with high number of visits from vulnerable populations did not have higher deaths. By focusing on population alone, we eliminate important factors such as state resources, advancements in practice of each states health care system, the socio-economic factors in the population, number of foreign tourists in the state, etc. New York has significantly more population and vulnerable population than Ohio, yet Ohio has more deaths from the flu the same is true for Illinois and Pennsylvania. More analysis is needed for staffing projections.
It is clear the number of deaths per area needs to be the center of concentration for further analysis. Other factors such as technological innovations, staffing, access to health care, best practices in each state, and access to preventative measures will all have some influence or correlation as population did, but more investigation into what is significantly important and key to determining factors in areas or weeks with increased deaths will need to be looked into further. Though some data analyzed was in surveys and is subject to human error, the analysis was conducted on the basis that data provided is accurate. Staffing ratios will need to be examined as well as states with higher outcomes and their practices. If the companies previous staffing records were available, staffing numbers in relation to deaths could be adjusted. Since this number is not available number of deaths will become the center of focus and population may still be used to help guide staffing recommendations with adjustments for areas with greater death rates, decreased access to health care, and older technological practices. Data will need to be obtained in these areas and analyzed to provide a more complete and accurate analysis to be used for guidelines for staffing recommendations.
Final presentation will include written report available to stake holders and power point presentation available to anyone attending board meetings to determine staffing assignments for Med Staffing- R-Us and areas where company should focus at gaining more contracts to help increase better out comes and decrease mortality rates. Final analysis will be available at Monthly board meeting conducted at the first of the month to help facilitate staffing assignments including long and short term contracts.
Current visual illustrate peak flu season trends and peak areas of need. With new analysis visuals will need to be added and revised for recommendations. Current visual trends can be viewed in Appendix for better understanding of analysis conducted thus far.
To view the original documents in Stephanie’s portfolio, head to the following links: