🏙️ Louisville Payroll Analytics – Part 4

📊 From SQL Results to Visual Story

With the SQL layer stable (tables, indexes, views), it was finally time to switch hats and think like a BI developer.

The goal wasn’t to produce 20 random charts, but a small dashboard that a non‑technical stakeholder could actually use.

I decided on four core views:

1. Yearly Trend: headcount and overtime % over time
2. Salary by Department (2024): total payroll and average salary
3. Overtime by Department (2024): cost and intensity of overtime
4. Salary Variance by Job Title (2024): internal pay ranges

---

🧩 Connecting Tableau to PostgreSQL

Steps:

• Data → PostgreSQL → connect to louisville_payroll
• Use the analytical views (v_department_summary, v_overtime_analysis, v_top_earners) instead of raw tables
• Verify field types (numeric vs string vs date)
• Create a data source per thematic area (trend, department metrics, job title metrics)

Working directly on the views kept Tableau calculations clean and pushed heavy work down to SQL, where it belongs.

---

📈 View 1 – Yearly Trend

A dual‑axis chart:

• Area: number of employees over time
• Line: overtime as a percentage of total payroll

This immediately shows if overtime is growing faster than headcount, which is a red flag in any organization.

---

🏢 View 2 – Salary by Department (2024)

Horizontal bar chart:

• X‑axis: total payroll
• Y‑axis: department (sorted descending)

Tooltips include:

• employee count
• average salary
• share of total payroll

This view answers “who costs the most?” at a glance, without touching SQL.

---

🔥 View 3 – Overtime by Department (2024)

Same structure as salary view, but:

• Bars = total overtime cost
• Color = overtime % of total payroll

This makes it impossible to ignore departments where overtime is both large in absolute terms and heavy as a share of payroll.

---

⚖️ View 4 – Salary Variance / Inequality

For this, I used a derived dataset (like Salary-Variance-per-Job-Title), built from SQL:

• Bars show salary range (max − min) per job title
• Filter to job titles with at least N employees
• Tooltip shows min, max, mean and std dev

This is the visual counterpart of the SQL analysis on internal equity.

---

📌 Dashboard Layout

The final dashboard layout:

The work is directly tied to immediate and simple questions:

How big is the payroll?
Where is overtime concentrated? Which roles have the largest internal gaps?

---

🎯 Next Steps

Link it with the GitHub repo in the README.
Write a final “retrospective” post about what worked and what didn’t.

To be continued…