Human-centered HR leadership, translated into measurable business impact.

Completed project · Fictional demonstration data

A fictional growing company, 2018 to 2025: 579 employees across 8 departments, 254 exits with type and reason, and 1,491 annual performance ratings. The data is stored in SQLite and every figure below comes from SQL queries.

Attrition by department, 2025

Leavers divided by average headcount (start and end of year). Under each team: leavers of average headcount.

Needs actionAbove company rateBelow company rateCompany rate

Key findings

1. Customer Support has a retention problem. It lost 29 people against an average headcount of 60 in 2025, an attrition rate of 48.3%. That is more than double the company rate of 20.1%, about double Sales (23.1%) and four times Engineering (11.7%).
2. New hires are leaving early, and it is getting worse. 23.5% of 2023 hires left within their first year, rising to 27.1% of 2024 hires. Roughly one in four new starters is gone within twelve months.
3. People leave for better opportunities and pay. “Better opportunity” and “Compensation” are the top two voluntary exit reasons in most departments. In Customer Support they account for about half of all voluntary exits (28.9% and 22.4%), with “Career growth” third (19.7%).
4. Leavers were rated lower. People who left averaged 2.73 out of 5 in performance reviews, against 2.99 for current staff. Lower ratings may signal poor fit, or a lack of support and development before people go, so this is worth investigating before drawing conclusions.

Recommendations

• Benchmark Customer Support pay. With compensation and better offers driving exits, compare Support salaries with the market and consider a targeted adjustment or retention incentive.
• Build a visible career path in Support. Define progression steps (agent, senior agent, team lead) with clear criteria, so people can see a future without leaving.
• Strengthen the first year. Introduce structured onboarding, 30/60/90-day check-ins and a buddy system, and track first-year attrition each quarter.
• Use exit and stay interviews. Confirm the reasons behind the numbers and reach at-risk employees before they resign.

How the analysis works

1. Calculate 2025 headcount at the start and end of the year for each department, then divide leavers by the average.
2. Follow each 2023 and 2024 hiring group for twelve months to measure first-year attrition.
3. Rank voluntary exit reasons within each department and show each reason’s share.
4. Compare average performance ratings of leavers with current staff.
5. Check the data for missing values and stray spaces before reporting.

WITH people AS (
  SELECT employees.department_id, hire_date, termination_date
  FROM employees
  LEFT JOIN terminations ON terminations.employee_id = employees.employee_id
),
by_dept AS (
  SELECT department_id,
    SUM(CASE WHEN hire_date <  '2025-01-01' AND (termination_date IS NULL OR termination_date >= '2025-01-01')
             THEN 1 ELSE 0 END) AS hc_start,
    SUM(CASE WHEN hire_date <= '2025-12-31' AND (termination_date IS NULL OR termination_date >  '2025-12-31')
             THEN 1 ELSE 0 END) AS hc_end,
    SUM(CASE WHEN termination_date BETWEEN '2025-01-01' AND '2025-12-31' THEN 1 ELSE 0 END) AS leavers
  FROM people
  GROUP BY department_id
)
SELECT department_name,
       (hc_start + hc_end) / 2.0 AS avg_headcount,
       leavers,
       ROUND(100.0 * leavers / ((hc_start + hc_end) / 2.0), 1) AS attrition_pct
FROM by_dept
JOIN departments ON departments.department_id = by_dept.department_id
WHERE hc_start + hc_end > 0
ORDER BY attrition_pct DESC;

What this demonstrates

SQL joins (including LEFT and self joins), aggregation with GROUP BY and HAVING, CASE logic, subqueries, CTEs, window functions and date calculations, applied to core HR measures: headcount, attrition rate, first-year attrition and tenure.

Independent portfolio demonstration using fictional data. Ratings are compared across all reviews on record. Differences between groups are observed patterns, not proof of cause.