An exploratory and statistical analysis of roughly 41,700 Airbnb listings across Europe — pricing, room type, location and value-for-money patterns.
Overview
An academic exploratory-analysis project covering approximately 41,714 Airbnb listings across Europe, examining pricing, room type, distance-from-center and value-for-money patterns.
Problem
Raw listing data needed real cleaning — missing values, duplicates, invalid observations and extreme price outliers — before any pricing pattern could be trusted.
Approach
Missing-value and duplicate analysis, median/mode imputation, invalid-observation removal, and price-outlier handling using the 99th percentile rather than a fixed cutoff. Feature engineering added Price per Person, Above City Average Price, Distance Category and Value for Money.
Engineering
- Missing-value analysis, duplicate checks and statistical summaries
- Price, room-type, distance and correlation analysis
- 99th-percentile-based price outlier handling
- Engineered features: Price per Person, Above City Average Price, Distance Category, Value for Money
Result / Current State
Completed as a university project and received an excellent grade. Key findings: entire homes tend to be more expensive than private or shared rooms, central listings command higher prices, prices are right-skewed, and meaningful outliers exist throughout the dataset.
What I Learned
Separating genuine premium listings from data-entry outliers required percentile-based rather than arbitrary thresholding, and cross-checking findings (like the entire-home price premium) against room type and location together rather than in isolation.
University project — excellent grade
