The Seattle Airbnb dataset is a comprehensive collection of data related to Airbnb listings in Seattle. It provides detailed information across several aspects of the Airbnb ecosystem, including listings, calendar availability, and user reviews.
The Seattle Airbnb dataset offers a rich source of information for analyzing various aspects of Airbnb listings in Seattle, including trends in pricing and availability, geographic distribution of listings, and customer satisfaction based on reviews. This dataset is valuable for conducting detailed exploratory data analysis (EDA), machine learning projects, and deriving insights into the short-term rental market in Seattle.
The dataset is divided into three main files:
listings.csv: Contains detailed information about each Airbnb listing in Seattle.calendar.csv: Provides daily availability and pricing information for each listing.reviews.csv: Contains user reviews for each listing.
(base)$: git clone git@github.com:mafda/seattle_airbnb_data_analysis.git
(base)$: cd seattle_airbnb_data_analysis-
Create the conda environment
(base)$: conda env create -f environment.yml
-
Activate the environment
(base)$: conda activate seattle_airbnb
-
Download the dataset from Seattle Airbnb Open Data, create
datafolder and copy the data here.(seattle_airbnb)$: mkdir data
.
├── README.md
├── data
│ ├── calendar.csv
│ ├── listings.csv
│ └── reviews.csv
├── environment.yml
└── src
└── seattle_airbnb_data_analysis.ipynbThis project will follow CRoss Industry Standard Process for Data Mining - CRISP-DM methodology.
-
Business Understanding: Identify the key business questions, such as factors influencing occupancy rates, pricing strategies, and customer satisfaction.
-
Data Understanding: Explore the structure and content of the dataset, including listings.csv, calendar.csv, and reviews.csv.
-
Data Preparation: Clean and pre-process the data by handling missing values and converting data types.
-
Modeling: Apply statistical and machine learning models to analyze the linear relationship between features and price. Use visualization techniques to illustrate the relationships.
-
Evaluation: Validate the models and analysis results for accuracy and reliability.
-
Deployment: Document the insights and recommendations based on the analysis. Provide actionable strategies for optimizing listings, pricing, and improving customer experience.
Some exploratory analysis.
The analysis reveals a relationship between occupancy rates and prices. This indicates a seasonal trend with peaks in winter (January) and summer (July), followed by gradual decreases. This correlation suggests that higher occupancy rates, likely driven by seasonal demand, are associated with higher prices.
The correlation matrix shows that there is a high correlation between price and review-related variables (such as number of reviews and review score). Significant correlations are also observed with the number of bathrooms, beds, bedrooms, accommodation capacity and number of guests included. This suggests that these factors are key determinants of listing prices, with larger, better-reviewed properties tending to command higher prices.
The most frequently used words in user reviews were "great," "clean," "location," and "comfortable." These keywords suggest that customers generally had positive experiences and appreciated the cleanliness, location, and comfort of the accommodations. The analysis shows that customer satisfaction is overwhelmingly positive, with a few areas for improvement in handling multilingual reviews.
Linear Regression is a simple model that attempts to capture the linear relationship between features and price.
- Pros: Easy to interpret and quick to train.
- Cons: May not capture complex relationships between variables.
Although the model has a reasonable level of predictive ability, an R² of 0.558 also suggests that there is 44.2% of the variability that is not being captured by the model, indicating that some important features could be missing or that the model is not complex enough to capture all relationships present in the data.
| Cost | Result |
|---|---|
| Mean Absolute Error (MAE) | 38.09 |
| R² Score | 0.558 |
Although the model has a reasonable level of predictive ability, an R² of 0.558 also suggests that there is 44.2% of the variability that is not being captured by the model, indicating that some important features could be missing or that the model is not complex enough to capture all relationships present in the data.
For a detailed analysis of the results obtained, visit our Medium blog. Here you will find a comprehensive explanation of the findings and their interpretation in the context of the Seattle Airbnb data analysis.
Seattle Airbnb: Temporal Trends, Price Influences, and Customer Satisfaction Analysis.
made with 💙 by mafda