Skip to content
Case Study · Public Health & Resource Planning

Influenza mortality mapped for resource planning

An end-to-end Excel and Tableau analytics project that turned 700K+ CDC flu-death records (2009-2017) into an eight view interactive Tableau story: age group risk, state level concentration, seasonal peaks, and a regional forecast; telling a medical staffing agency where, when, and for whom to plan ahead.

Excel Tableau Data Cleaning EDA Forecasting Public Health
Project overview

From 700K CDC death records to a seasonal staffing plan

The project was built around a practical operational question: where is flu mortality highest, when do seasonal peaks occur, and which age groups need the most resource-planning attention?

Business problem

A medical staffing agency supports hospitals across all 50 states and must decide when and where to send temporary staff during flu season, particularly for vulnerable populations. Raw CDC mortality data is not decision ready. The goal was to turn it into clear, data driven guidance on the timing and geographic distribution of staffing needs.

Analytical goal: move beyond raw death counts to a population-normalised view of senior risk that is fair to compare across states of very different sizes.

Final output

The final deliverable was an eight view interactive Tableau story fed by a cleaned validated Excel dataset (CDC mortality) joined to US Census population, covering national totals, age and seasonality, state concentration, senior specific risk, a regional forecast, and a findings with next steps summary.

The story helps stakeholders see which states carry the highest burden, where seniors are most at risk per capita, the recurring November&-;January peak, and a clear set of staffing priorities.

Deaths analysed700K+CDC influenza mortality records, 2009-2017
65+ deaths412Kthe single largest age group with the majority of all flu deaths
Tableau views8national overview through to forecast and next steps
Datasets merged2CDC mortality + US Census population, joined for per-capita rates
Excel-first workflow

How the analysis was built

The Excel work turned raw, inconsistent government health data into a profiled, validated, analysis-ready dataset, the foundation everything visual rests on, before a single chart was drawn in Tableau.

01
Import & reviewImported the raw CDC influenza mortality data and reviewed it for missing, duplicated, and inconsistent entries before any analysis.
02
Data profilingDocumented every variable by data type, measurement level (nominal / ordinal / discrete / continuous), time property, and structure; a transparent profile of the raw data.
03
Integrity assessmentCalculated MIN, MAX and AVERAGE descriptive stats and built frequency tables for state, year, gender and cause to flag outliers, spelling variations, and non-standard names.
04
CleaningStandardised state and region names, reformatted dates for time-series use, removed redundant fields, applied uniform numeric precision, and added missing category labels.
05
Re-profile & validateRe-profiled the cleaned data, recalculated summary statistics, and built a separate integrity check confirming each variable’s min, max and mean now fell in realistic ranges.
06
Census datasetRepeated the full profiling and cleaning process on the US Census population data used for normalisation and rate calculations.
07
MergeJoined mortality and population data with VLOOKUP to enable population normalised mortality rates; comparing risk, not just raw counts.
08
Tableau buildImported the final dataset and built an eight-view interactive story: line, bar, pie, scatter, choropleth maps, and a regional forecast with year and region filters.
09
Findings & recommendationsDocumented insights and translated them into staffing and preparedness recommendations for a non-technical public-health audience.
The method

Per-capita senior risk, not just raw death counts

Large states will always post the biggest totals, so a raw count is misleading across states of very different sizes. Normalising each state’s senior deaths against its senior population makes the comparison fair and defensible.

How the senior-risk view works

The map shades each state by the share of flu deaths among people aged 65+, not the raw number, so population size does not dominate the picture, and it never claims a clinical or causal finding.

01Total flu deaths set the overall burden picture (state background colour, 6.4K-57.8K).
02A 65+ death-count layer shows where seniors are most affected in absolute terms (sized circles).
03A population-normalised “% senior death” measure reveals per-capita risk (0.05 to 0.36 across states).
04A correlation view tests senior population against senior deaths, region by region.
Highest 65+ deathsCalifornia 47.5K

The single largest senior flu death total 2009-2017, ahead of New York (36.6K) and Texas (22.1K).

Highest per-capita riskAlaska, Wyoming, Vermont

Darkest on the % senior-death map, a high share of senior deaths even where total numbers are modest.

Strongest correlationWest & Northeast

Senior population and senior deaths track most tightly in these regions, pointing to population-driven demand.

Two views, by designVolume ≠ risk

The biggest total is not always the highest risk, a smaller state can need proportionally more support.

Tableau influenza mortality and resource-planning interactive dashboard preview
The data at a glance

The data at a glance

Headline cuts of the 700K+ death records, all drawn from the cleaned CDC and US Census datasets.

Deaths by age group

Total flu deaths 2009-2017.

  • 65+ Years412K
  • 35&-;64 Years108K
  • 05&-;34 Years83K
  • 0&-;4 Years55K

Worst years

Highest annual death totals.

  • 201579.5K
  • 201379.2K
  • 201777.9K
  • 201176.3K

Peak season

Deaths by month; winter-dominant.

  • January76.9K
  • March66.1K
  • December63.7K
  • October52.6K

Top states (65+)

Senior flu deaths by state.

  • California47.5K
  • New York36.6K
  • Texas22.1K
  • Pennsylvania20.6K
Tableau story

An eight-view story built around stakeholder questions

The story moves from a national overview into seasonality, geography, senior-specific risk, and a forward-looking forecast with recommendations.

01

National Overview

KPI cards by age group and the human cost framing; who is most at risk from flu, and where, across all 50 states.

02

Seasonality & Age

Yearly and monthly death trends, the age-group breakdown, and a senior population-vs-deaths correlation scatter.

03

Geography & Senior Risk

A total-deaths state map, the top-10 states ranking, a compare-your-states tool, and the per-capita % senior-death map.

04

Forecast & Next Steps

A regional forecast with prediction bands, plus prioritised findings, recommendations, and next steps.

Key findings

What the data made clear

The story turns 700K+ death records into a small number of decisions a public-health staffing team can actually act on.

Seniors carry the overwhelming majority

Adults aged 65+ accounted for 412K of the flu deaths more than every other age group combined, making them the clear priority for staffing and vaccination outreach.

Deaths peak every winter, with a March echo

Mortality consistently peaks November-January (January highest at 76.9K), followed by a notable secondary wave in March (66.1K), a recurring pattern staffing can be planned around.

Volume and per-capita risk are different stories

California, New York and Texas post the biggest totals, but Alaska, Wyoming and Vermont carry the highest share of senior deaths, so the busiest states are not always the most at risk per capita.

The pattern is predictable, not random

Years 2013, 2015 and 2017 were the worst, and a regional forecast projects continued seasonal recurrence every winter, confirming flu demand can be planned for, not just reacted to.

Technical validation

Cleaned, validated, and honest about its limits

A public-health dashboard only holds up if the cleaning is documented, the totals are checked, and the analysis is clear about what it can and cannot claim. All three were built in from the start.

Numbers you can defend

The dataset was not taken on trust; every transformation was documented and re validated before any chart was built.

Documented cleaning: every fix: standardised state names, reformatted dates, uniform numeric precision, recorded in a dedicated “Data Profile of Cleaned Data” sheet.
Re-validated integrity: a separate integrity check confirmed each variable’s min, max and mean fell in realistic ranges after cleaning, with no invalid categories left.
Honest forecast: the projection uses Tableau’s exponential-smoothing forecast with confidence bands, presented as a trend-based projection, never a precise prediction.
Cleaned, profiled datasets

CDC mortality and US Census population data, each profiled, cleaned, and re-validated in Excel before use.

Population-normalised risk measure

A % senior-death metric separating per-capita risk from raw counts across all states.

Eight-view Tableau story

National overview, seasonality, age, geography, senior risk, correlation, forecast, and recommendations.

Staffing recommendations

Prioritised, plain-language actions tied to peak months, high-risk states, and vulnerable populations.

Recommendations

How a staffing agency could use the dashboard

Staff up for the winter peak: concentrate temporary staffing in high-risk states across the November&-;January peak, and account for the smaller March second wave.

Protect vulnerable seniors first: allocate extra support where the senior population is especially at risk per capita, even where total numbers are smaller.

Plan from the forecast: use the regional forecast and year-to-year patterns to anticipate surges rather than react to them.

Know the limits: the analysis guides the timing and geography of staffing; it deliberately stops short of clinical or hospital capacity claims the data cannot support.

Tools & deliverables

What this project demonstrates

End-to-end data profiling and cleaning of messy government data, descriptive statistics and frequency analysis for quality assurance, population normalisation, multi-view Tableau storytelling, trend forecasting, and the ability to turn a large public-health dataset into prioritised, defensible recommendations.

Excel Tableau Data Cleaning Data Profiling EDA Descriptive Statistics Forecasting Data Storytelling

Need the full portfolio or resume?

Download the PDF portfolio for a polished overview of the projects, or open the resume for the formal career summary, tools, and work history.