Mod 3 Project – Theme Park Analytics

By Debo Odutola

📂 Project Overview

This project analyzes a theme park’s guest behavior using a star-schema database (themepark.db).
The goal was to prepare clean data, run exploratory SQL, and generate Python visuals to provide actionable insights for three stakeholder groups: General Manager (GM), Operations, and Marketing.

---

🛠️ Process Summary

1. Schema Review

   • Verified star schema structure (dim_guest, dim_ticket, dim_attraction, fact_visits, fact_ride_events, fact_purchases).
   • Confirmed primary and foreign key integrity (no orphaned keys).

2. Data Cleaning

   • Removed duplicates from fact tables (fact_ride_events) while preserving one record.
   • Standardized categorical values:
     • dim_guest.home_state → collapsed variants (NY, NEW YORK → NY).
     • fact_purchases.payment_method → standardized (cash, card, CARD,, Apple Pay → CASH, CARD, APPLE PAY).
     • dim_attraction.attraction_name → trimmed punctuation/case (Galaxy Coaster vs Galaxy coaster; Pirate Splash! vs Pirate Splash).
   • Normalized casing and trimmed whitespace in string columns.
   • Validated numeric fields (wait times, spend, ratings) and set out-of-range values to NULL.

3. Exploratory SQL Queries (EDA)

   • Attendance trends over time.
   • Guest demographics and ticket breakdown.
   • Ride utilization and satisfaction.
   • Purchases by category and payment method.

4. Note

   • While building Python visuals, I noticed data quality issues (e.g., duplicate state names, inconsistent payment methods).
   • This sent me back to SQL to improve cleaning.
   • Visualizations are a powerful tool for catching hidden inconsistencies.

---

📊 Python Visuals

1. Daily Attendance & Spend

• Attendance and revenue rise together, indicating strong per-guest spend consistency.

2. CLV by State

• Certain states (NY, CA, TX) contribute disproportionately to guest lifetime value.
• Useful for targeted marketing campaigns and loyalty programs.

3. Attraction Wait Time vs Satisfaction

• Longer waits (>30 minutes) correlate with lower satisfaction.
• Popular thrill rides (coasters, water rides) suffer the steepest drops when waits exceed 45 minutes.
• Operations should implement queue management or capacity redistribution.

---

💡 Insights & Recommendations

General Manager (GM)

• Plan staffing and hours around peak holidays (attendance & revenue spikes).
• Balance growth (more guests) with experience (avoid long waits).

Operations

• Reallocate staff to high-wait attractions.
• Introduce virtual queues or staggered openings to protect satisfaction.

Marketing

• Focus campaigns on high-value states (NY, CA, TX).
• Promote in-park purchases, not just admissions, to increase per-guest revenue.
• Leverage CLV segmentation for repeat-guest loyalty programs.

---

⚖️ Ethics & Bias

• Data cleaning ensured fairness (e.g., resolving duplicate states so one region isn’t over/underrepresented).
• Analysis is limited to in-park behavior — we avoid inferring demographics or socioeconomic status beyond the data.
• Recognize that high-spending states may correlate with travel privilege, so recommendations should balance inclusivity.