Gonen8/Ski_Resorts_World_Wide

Ski Resorts Worldwide - Exploratory Data Analysis (EDA)

Author: Omer Gonen Course: Introduction to Data Science
Assignment: EDA & Dataset Analysis
Dataset Source: Kaggle – Ski Resorts Worldwide Dataset
Dataset Size: ~5,478 rows × 33 columns

---

Project Goal

The goal of this project is to explore ski resorts worldwide to understand how physical terrain characteristics influence resort quality ratings. This analysis investigates the relationship between measurable resort features—such as altitude, trail size, and infrastructure—and user satisfaction ratings to identify the key physical factors that contribute to high-quality ski resort experiences.

---

Dataset Overview

The dataset contains information about ski resorts worldwide with the following key columns:

• NameResort – Name of the ski resort
• Country – Country where the resort is located
• Continent – Geographic continent
• Altitude – Elevation of the resort (meters)
• Total Kms – Total kilometers of ski trails
• Easy – Number of easy trails
• Intermediate – Number of intermediate trails
• Difficult – Number of difficult trails
• Ski_resort_size – Size category (1-5)
• Lifts_cable_cars – Number of lifts and cable cars
• Rate – Resort quality rating (0-5 scale)

The data was cleaned, validated, and filtered to ensure analysis quality.

---

Data Cleaning Summary

• Selected 11 relevant columns from the original 33 columns for analysis.
• Removed 22 irrelevant columns including: Mountain restaurants, ski huts, gastronomy, Families and children, Après-ski, website URLs, adult-only indicators, summer skiing availability, glacier presence, snow parks, night skiing, and other non-essential features...
• Note: Some potentially relevant columns (e.g., Snow reliability) were excluded due to insufficient data availability.
• Fixed column names by removing trailing spaces and replacing spaces with underscores (e.g., Ski resort size → Ski_resort_size, Lifts and cable cars → Lifts_cable_cars).
• Removed resorts with Total Kms = 0 (no skiable terrain).
• Removed resorts where all trail types = 0 (Easy, Intermediate, Difficult all equal zero).
• Retained only resorts with valid rating data (Rate not null).
• Cleaned Lifts_cable_cars column by converting string values with commas to numeric format (e.g., '5,0' → 5.0).
• Final dataset: 2,338 valid resorts for analysis (42.7% retention rate).

---

Research Questions & Insights

Q1: Which terrain characteristics show the strongest correlation with resort quality ratings?

Visualization: Correlation Heatmap of Key Variables

Click on the image above to view it in full size

Insight:
Total Kilometers demonstrates the strongest correlation with resort ratings (r = 0.690), followed by Lifts & Cable Cars (r = 0.682) and Altitude (r = 0.629).

This indicates that resort size is the most important predictor of resort quality. Larger resorts with more trail kilometers offer greater variety and options for skiers of all skill levels. Infrastructure (lifts) and elevation also play significant roles, with all three factors showing strong positive correlations with quality ratings.

---

Q2: Does resort altitude affect quality ratings? Is there a measurable relationship between elevation and resort performance?

Visualization: Average Altitude by Resort Rating

Click on the image above to view it in full size

Insight:
There is a clear, strong positive relationship between altitude and resort quality.

• Lowest-rated resorts (Rating ≤ 2.0): Average altitude ~1,300 meters
• Highest-rated resorts (Rating = 5.0): Average altitude ~1,900 meters
• Altitude difference: Approximately 600 meters (46% increase)

Higher-elevation resorts consistently receive better ratings. This substantial elevation difference translates to better snow quality, reliability, and overall skiing conditions. The correlation between altitude and rating is strong (r = 0.629).

---

Q3: How are high-quality ski resorts (rating > 3) geographically distributed across continents?

Visualization: High-Rated Resorts (Rating > 3) by Continent

Click on the image above to view it in full size

Insight:
Europe dominates the high-quality ski resort market, accounting for 79.1% of all top-rated resorts (393 out of 497).

• Europe: 393 resorts (79.1%)
• North America: 85 resorts (17.1%)
• Asia: 10 resorts (2.0%)
• Oceania: 5 resorts (1.0%)
• South America: 4 resorts (0.8%)

This geographic concentration suggests that European regions possess unique advantages for ski resort development—including Alpine geography, consistent snow conditions, established ski culture, and mature infrastructure.

---

Final Conclusions

Main Research Question:
"How do physical terrain characteristics—including altitude, total trail kilometers, and number of lifts and cable cars—affect ski resort quality ratings?"

Answer: YES—physical terrain characteristics significantly and measurably affect ski resort quality ratings.

Key Findings:

1. Resort size is the dominant quality predictor

   • Total kilometers show the strongest correlation with ratings (r = 0.690)
   • Larger resorts offer more variety and options for different skill levels
   • More trails = better experience and higher satisfaction

2. Infrastructure quality is critically important

   • Number of lifts shows very strong correlation (r = 0.682)
   • More lifts = better accessibility, reduced wait times, and improved experience
   • Modern lift systems are essential for high ratings

3. Altitude significantly affects quality

   • Elevation shows strong correlation with ratings (r = 0.629)
   • Top-rated resorts average 600m higher elevation
   • Higher altitude = better snow quality and longer seasons

4. Geographic concentration of quality

   • Europe leads with 79% of high-rated resorts
   • Location matters: certain regions have natural and infrastructural advantages

Overall Insights:

• All three physical terrain characteristics show strong positive correlations with quality
• Resort size (Total Kms) is slightly more important than infrastructure or altitude
• No trade-offs observed—all factors contribute positively to ratings
• Geographic distribution is far more concentrated than expected (Europe's dominance)
• Data-driven approach reveals clear, actionable patterns for resort development and investment

---

Video

[Video link placeholder]