airbnb-italy

Airbnb Market Analysis in Venice, Italy

by Luca Scarpantonio

---

Project Overview

This project explores how the Airbnb market in Venice has evolved over the past decade, focusing on changes in supply, pricing, and host behavior.
The analysis follows the CRISP-DM process — from data gathering and cleaning to modeling and visualization — and aims to answer a set of real-world business questions relevant to the tourism economy of Venice.

The notebook was developed as part of the Udacity Introduction to Data Science Nanodegree.

---

Project Objectives

The analysis investigates key questions about Airbnb’s role in Venice:

1. How has the number of Airbnb listings evolved in recent years?
   → Understanding the growth trend of Airbnb activity in Venice.

2. How have supply and demand affected listing prices over time?
   → Exploring the relationship between market expansion and average nightly rates.

3. How has the number of active hosts evolved compared to listings?
   → Identifying whether growth comes from new hosts or increased host activity.

4. Do price patterns vary across Venice’s geographical areas?
   → Analyzing spatial segmentation (mainland vs. central Venice vs. islands).

5. Can we predict listing prices based on room type and location?
   → Building a simple regression model to identify the most relevant price factors.

---

Dataset

• Source: Inside Airbnb
• City: Venice, Italy
• Files Used: listings.csv, neighbourhoods.csv
• Temporal coverage: 2015–2025
• Observation:
  Dates are based on review timestamps, assuming that review activity correlates with the period when the property was actively listed.

Data Wrangling Steps

1. Removed duplicate and inconsistent records.
2. Converted prices from string (€/$) to numeric format.
3. Filtered unrealistic prices (price < 1000).
4. Removed missing or non-informative values (Not Available, NaN).
5. Extracted temporal features from last_review (year, month).
6. Aggregated average price and listings per month/year.

---

🔍 Methodology — CRISP-DM Process

Phase	Description
1. Business Understanding	Framed analytical questions around market growth, price evolution, and host behavior.
2. Data Understanding	Loaded and inspected Airbnb Venice listings; assessed missing data and outliers.
3. Data Preparation	Cleaned and transformed data, derived temporal and spatial features.
4. Modeling	Built a linear regression model to predict prices based on location and room type.
5. Evaluation	Interpreted model coefficients and evaluated fit (R²).
6. Deployment	Communicated insights via visualizations and this notebook.

---

Key Findings

###🏙️ 1. Rapid Market Expansion

• The number of active listings grew exponentially after 2016.
• Despite the pandemic slowdown in 2020, Airbnb’s presence recovered quickly, reaching record highs by 2025.

2. Price Stability Despite Supply Growth

• Average nightly prices remained stable between €160–€200 even as listings increased.
• This indicates a mature and competitive market, suggesting accessibility for travelers.

3. Persistent Spatial Segmentation

• Central Venice (San Marco, Dorsoduro, Cannaregio) maintains premium prices.
• Mainland and island areas (Marghera, Lido, Giudecca) are significantly cheaper.

👥 4. Host Dynamics

• The number of active hosts also increased, but the average number of listings per host rose faster, suggesting professionalization among hosts.

5. Predictive Insights

• Linear regression identified room type and location as the strongest predictors of price.
• The model confirmed the price gap between central and peripheral areas.

---

Visual Highlights

Visualization	Insight
Line plot of listings per year	Airbnb expansion over time
Scatter plot (Listings vs. Price)	Relationship between market growth and affordability
FacetGrid by area	Spatial variation of prices
Regression model output	Price prediction by room type and neighbourhood

---

Conclusion

The Airbnb market in Venice evolved from an emerging platform to a mature, balanced ecosystem.
While central districts remain premium, the rise of listings in peripheral areas has democratized access to the city, making Venice more affordable for short-term visitors.

In essence:
Airbnb reshaped tourism in Venice, balancing exclusivity and accessibility — preserving the charm of the lagoon city while opening it to a wider audience.

---

Tools and Libraries

• Python 3.11
• pandas, numpy
• matplotlib, seaborn
• scikit-learn
• jupyter, nbconvert

---

---

Author & Acknowledgments

Author: Luca Scarpantonio
Nanodegree: Udacity — Introduction to Data Science
Dataset source: Inside Airbnb
Inspiration: Airbnb market studies in European cities

---

For Further Reading

This analysis is also available as a Medium-style article:

“How Airbnb Transformed Venice: A Data-Driven View on Tourism and Accessibility”

---

Key Takeaway for Portfolio Reviewers

This project demonstrates a complete end-to-end data science process: from data cleaning and visualization to modeling and business storytelling — all focused on a real-world dataset that reflects both technical and analytical expertise.

This site is open source. Improve this page.