Intro

After recently finishing a few confidential reports for clients interested in understanding more about their data and programs, I decided to make up some data and bring together some of the more interesting analysis and visualization methods from that work into a report that I could share publicly to demonstrate my skills.

For this demonstration, I pretend the client (“Acme Corp”) is an national organization and the use case is that they’re interested in analyzing how their employees are engaging with the company volunteer program. There are a number of different ways that employees can use the company-sponsored volunteer hours, and the company would like to understand how the program is being utilized.

Goal

This report demonstrates my abilities across the following:

  • Tailoring my analysis to a variety of angles, e.g. temporal, spatial, and aggregating data at different levels.
  • Communicating results clearly, both in writing and through effective data vizualizations.
  • Using R Markdown and additional css stylesheets and HTML to develop dynamic reports that are both beautiful and intuitive.

Setting The Stage

Hey Acme Corp, welcome to the Volunteer Program Report. This report examines data for 2023 and 2024 and only active employees in relation to their participation with the company volunteer program.

There are 8 distinct ways that employees could participate with the volunteer program:

Data Sources
Source Date Collected Description
All Employees 12/31/24 All active employees
Community Cleanup 12/31/24 Employees participate in local park and neighborhood cleanups to improve community spaces.
Mentorship Program 12/31/24 Volunteers mentor students or young professionals, offering career guidance and skill development.
Nonprofit Fundraising 12/31/24 Employees help organize and run fundraising events for nonprofit organizations.
STEM Education Workshops 12/31/24 Volunteers lead hands-on STEM workshops for students, promoting careers in science and technology.
Food Bank Support 12/31/24 Employees assist at food banks by sorting, packing, and distributing food to families in need.
Disaster Relief Volunteering 12/31/24 Volunteers provide emergency support in response to natural disasters, including aid distribution and rebuilding efforts.
Animal Shelter Assistance 12/31/24 Employees spend time at animal shelters, helping with pet care, adoption events, and facility maintenance.
Environmental Advocacy 12/31/24 Volunteers participate in environmental initiatives such as tree planting, advocacy campaigns, and sustainability education.

Overall Activity

Let’s start broadly with exploring the amount of volunteer activity — I’ll just shorten that to “activity” or “activities” from here on out — with Acme Corp regardless of department. There are 176 total employees. Of those, 85 participated in at least one Acme Corp activity at some point during the year, which together accounts for a total of 300 unique activities. On average, employees participated in 1.7 activities across the two years in the data, and the most engaged volunteer in the company participated in 10 activities.

However, 91 employees (52 percent!) did not volunteer in any way.

Summary by Department

This table breaks these activities down by type and department. Not too much to dig into here yet — it’s mostly nice for reference or for informing things like annual reporting.

Activity
Type		 Accounting Data Engineering Finance Marketing Sales Total
Animal Shelter Assistance 1 2 3 0 1 1 8
Community Cleanup 6 14 17 4 12 26 79
Disaster Relief Volunteering 1 5 9 3 8 11 37
Food Bank Support 7 3 16 4 7 18 55
Mentorship Program 5 4 7 3 4 4 27
Nonprofit Fundraising 4 3 6 3 7 9 32
STEM Education Workshops 3 3 3 4 3 7 23
Environmental Advocacy 0 2 14 1 8 14 39
Total 27 36 75 22 50 90 300
Total Employees 22 25 54 15 22 38 -
Activities per Employee 1.2 1.4 1.4 1.5 2.3 2.4 -

We can explore these total activities more deeply by comparing visually across the groups. The Sales department is buoyed by high activity among some key folks like [random name, random name, random name…]. They’re the second-largest department, yet produce 2.4 activities per employee — more than any other department.

However, at 2.3 activities per employee, the Marketing department is not far behind either — mostly thanks to lots of activity from [random name, random name, random name…]

Permutation Testing

The next logical question after seeing these differences in the activity rates is, great, but are those differences meaningful? It’s impossible to tell intuitively if the Sales department’s activities-per-employee rate of 2.4 is significant when compared to something like the Finance department’s rate of 1.4, or if that difference could just be due to chance.

We can apply statistical testing to this question using an approach called permutation testing. This method assesses whether an observed difference between departments is meaningful by randomly shuffling the data, recalculating the difference, and repeating this thousands of times. In our case, the random shuffling simulates a fake world where there is no association between the department and each person’s likelihood to participate. By comparing the original result to these random outcomes that behave under the assumption that there is no link between department and activity rate, we can see how unusual the original observed difference is relative to how common or unusual it would be when there is no association between those characteristics.

Let’s use the Sales department (highest per-employee activity rate) versus the Accounting department (lowest) as an example. I ran 5,000 permutations, and we see that that the observed result is highly uncommon among the simulations. This is strong evidence that the differences between these departments are not due to chance; there is some sort of meaningful difference in the propensity of employees to volunteer.

For context, here are the permutation results between the Finance and Accounting departments. Here we see that the observed result is much more common among the permutations; about 45% of the observations are just as extreme or more than it is. This overlap is technically a p-value (0.45), which lots of folks are at least semi-familiar with. With a p-value this large, the difference in volunteering rates between the Finance and Accounting departments teams is likely due to chance.

Leaderboards

Individuals

For a true across-the-business leaderboard, let’s look at the overall top 20 by raw number of volunteer activities:

Departments

We can also break activity down by individuals across departments.

Volunteering by Type and Department

And we can look at what types of volunteering are most popular for each department with these handy radar charts.

Each point represents how many volunteer activities employees in a department have logged across the eight different activity types. The Sales department, for instance, have logged 26 Community Cleanup activities — the most of any activity for a single department — so their point for that activity is all the way out at the maximum line of 26.

Sales

Marketing

Data

Engineering

Accounting

Finance

Departmental Volunteering %

Another way to look at the activity across departments is by breaking down the percentage of employees in each department who either participated in at least one activity versus those who did not participate at all.

Activity by Seniority

I was also curious if the seniority was associated with different levels of overall volunteerism. However, there don’t appear to be notable differences in the volunteer rates among the different types of employees.

The Non-Participants

You can explore the specific people within these non-participating employees in the Activity by Department section.

Spatial Analysis

Activity by State

Interactive map of volunteer activity by employee state. Hover to see totals.

Timeline Analysis

Annual Timeline

This chart includes a smoothed average trend line, calculated using nearby data points to create a locally weighted curve that highlights overall trends while reducing short-term fluctuations.

Quarterly Averages

One way to think about tracking data across time is to break the volunteerism down into averages by employee for each quarter. If you remember from the beginning of the report, there are 176 employees in Acme Corp who cumulatively participated in 300 activities. On a per-employee per-quarter basis, that breaks down to:

(total activities / total employees) / 8

or 0.21 average activities per person each quarter for the Acme Corp as a whole.
Acme Corp Quarterly Average 0.22

This gives us a reference point to assess the per-employee activity rate for each group across 2024. The data are normalized to allow comparison across groups of different sizes. The table cells below are color-coded in reference to the overall Acme Corp quarterly average: darker greens are farther above the overall average, darker purples are farther below, and the whites are nearly identical.

This table provides two main benefits:

  • You can see differences across quarters. 2024 had particularly strong volunteer rates within Sales and Marketing, for instance.
  • It makes temporal trends within a single team visible, especially as future data get added. This will allow Acme Corp to see how the per-member volunteer rate within a particular department changes over time.
Quarterly Per-Employee Average Activities
Department
2023
2024
 Average
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
Sales 0.32 0.18 0.24 0.37 0.34 0.11 0.32 0.50 0.30
Marketing 0.45 0.18 0.14 0.32 0.41 0.32 0.36 0.09 0.28
Finance 0.20 0.00 0.27 0.33 0.20 0.20 0.20 0.07 0.18
Data 0.20 0.40 0.08 0.08 0.24 0.12 0.04 0.28 0.18
Engineering 0.17 0.19 0.13 0.19 0.20 0.17 0.11 0.24 0.17
Accounting 0.18 0.09 0.18 0.09 0.14 0.14 0.14 0.27 0.15

Individual Activity Breakdown

A visual breakdown of activity counts by department, with each employee colored according to their investment level. All charts are set to the same x-axis to allow for easy comparison across departments.

Remember that you can refer back to the Summary by Department section to see the per-department breakdown of total employees, activity types, and total activities.

Marketing

Engineering

Accounting

Finance

Sales

Data

Recommendations

[Note]: for this demonstration I’ve stripped out this recommendations section as it doesn’t make a whole lot of sense to write up a set of made-up recommendations for a bunch of made-up data! But in the real versions of this general type of report this section has included a detailed list of recommendations for improving program buy-in and efficacy based on the analysis above and my knowledge of the client’s business.