Airbnb Market Analysis
This project explores short-term rental trends using real data from Inside Airbnb. The goal is to uncover pricing patterns, revenue dynamics, guest sentiment, and neighborhood trends to better understand what drives performance on the Airbnb platform.
Project Overview
Purpose
- Collect, clean, and analyze large real-world datasets
- Use SQL for complex business logic and performance analysis
- Apply Python for sentiment analysis and text mining
- Deliver actionable insights from raw data
- Combine technical skills with business intuition
Tech Stack
- PostgreSQL via pgAdmin4
- Python (pandas, TextBlob, SQLAlchemy)
- Jupyter Notebooks
- Markdown + SQL scripts
Key Questions Answered
- Which neighborhoods and room types dominate the market?
- How do prices vary by neighborhood and room type?
- What drives revenue and review engagement?
- How has activity changed over time?
- What are the sentiment trends from guest reviews?
Visualizations & Insights
1. Average Price by Room Type and Neighbourhood
Note: The unusually high average prices in some neighborhoods are due to extreme outliers. These have not been excluded from the current visualization but are noted for context.
- Entire home/apartment listings dominate the market and carry the highest average nightly price. Shared rooms are rare and priced significantly lower.
- SoHo and Williamsburg offer a strong balance of availability, guest satisfaction, and pricing power — signaling market strength.
2. Market Activity Projection (into 2026)
- Listing activity remains strong throughout the year with predictable dips in winter.
- Midtown Manhattan continues to see strong booking behavior despite premium pricing — a magnet for both tourists and professionals.
3. Average Airbnb Listing Price by Neighbourhood
- Brooklyn offers affordability with high listing volume, while Manhattan commands higher average prices.
- A few high-performing hosts account for disproportionate revenue — reinforcing market concentration.
Additional Insights
Revenue and Pricing Trends
- Listings with low availability (high bookings) significantly outperform others in terms of estimated annual revenue.
- The top-earning listings can bring in over $100,000/year, primarily those that are both expensive and nearly fully booked.
- Superhosts charge, on average, ~$25 more per night than non-superhosts. However, they also receive more reviews per month, suggesting guests are willing to pay a premium for reliability and service.
Booking Behavior and Activity
- Listings with more reviews per month are strong proxies for consistently booked listings and often align with higher revenue earners.
Guest Sentiment Analysis
Explore Full Sentiment Analysis Report (HTML)
- The average sentiment across all reviews was highly positive (~0.75 on a scale from -1 to 1), reinforcing the strength of the NYC Airbnb experience.
- SoHo, Park Slope, and Greenpoint had the most positively reviewed listings, while JFK-adjacent neighborhoods and Harlem had more negative feedback related to noise, cleanliness, or host communication.
- Keywords in positive reviews often included: “clean,” “responsive host,” “great location,” and “comfortable bed.”
- Negative reviews flagged recurring issues like: “dirty,” “noisy,” “bugs,” and “cancellation.”
- Listings with high sentiment and high review count are rare — but they’re ideal for identifying top-tier performers.
Additional Insights
- Some listings are priced very high but underperform in sentiment or review count — potentially indicating poor value perception.
- The most reviewed listing in the dataset had over 600 reviews, significantly above average, and priced competitively in the $90–$120 range.
How to Reproduce
- Clone the repo:
git clone https://github.com/SebastianMarrero/AirBNBAnalysis.git - Install required Python packages
- Set up PostgreSQL and import datasets
- Run the notebook:
jupyter notebook notebooks/airbnb_analysis.ipynb
For full documentation, see the project's README on GitHub.