A medical staffing agency seeks to provide temporary staffing to hospitals across the US for the upcoming influenza season. I analyzed two datasets to figure out when to send the staff and what amount to send to each state. The goal is to prevent understaffing or overstaffing as much as possible.
Deliverables
Provide information to support a staffing plan, detailing what data can help inform the timing and spatial distribution of medical personnel throughout the United States.
Determine whether influenza occurs seasonally or throughout the entire year. If seasonal, does it start and end at the same time (month) in every state?
Prioritize states with large vulnerable populations. Consider categorizing each state as low, medium, or high-need based on its vulnerable population count.
Assess data limitations that may prevent you from conducting your desired analyses
Tools
Excel: Data cleaning, integration, transformation, vlookup, hypothesis testing, visual analysis, forecasting
Tableau: Storytelling in Tableau, Presenting results
DatasetsCDC Influenza deaths by geography, time, age, and gender US Census Bureau Population data by geography
Business understanding: I read through the project brief to understand what was needed and came up with some questions I would answer to meet the business goals. I then designed my research project and formulated a hypothesis around the assumption that vulnerable populations were more likely to die from influenza.
Vulnerable populations are patients likely to develop flu complications requiring additional care, as identified by the Centers for Disease Control and Prevention (CDC). These include adults over 65 years, children under 5 years, and pregnant women, as well as individuals with HIV/AIDs, cancer, heart disease, stroke, diabetes, asthma, and children with neurological disorders.
Data understanding: I looked at a wide selection of datasets and decided to work with the CDC datasets and the US Census data sets since they contained enough information and were least likely to have bias. I then identified any other potential biases and limitations. I then generated data profiles for each dataset to understand the contents and recorded everything in a report.
Data preparation: At this stage, I performed data quality checks, cleaned each dataset, and performed descriptive analysis before integrating both datasets.
Analysis: I began with statistical analysis where I tested my hypothesis, I included my findings in an interim report. I then went on to perform spatial analysis and visualize my results in Tableau.
Presentation: I put together a Tableau storyboard and presented it to stakeholders.
Hypothesis: Vulnerable populations are more likely to die from influenza than non-vulnerable populations.
I tested this hypothesis with a two-tailed t-test and the results confirmed my assumption.
I could now base the distribution of temporary staff on the number of vulnerable populations in each state. States with higher vulnerable populations than other states will get more temporary staffing.
Geographic visualization showed a positive correlation between the total population and the vulnerable population in each state.
This meant I could now base the distribution on the total population in each state. Highly populated states like California would receive more temporary staffing than less populated states like Wyoming.
After further visualization, I discovered that less populated states had higher influenza death rates than more populated states.
This could be due to the company neglecting these states because of their low vulnerable populations.
I did not have enough data to confirm this so I reported this to the stakeholders. I suggested retrieving feedback from previous years to figure out how to properly staff the less populated states.
The influenza season occurs from December to March, so staffing should be provided during these months.
States with high populations should be prioritized.
States with low populations should not be neglected.
Feedback from previous influenza seasons could aid in making more informed decisions.
Interim reportTableau PresentationVideo presentation to stakeholders