Boston Airbnb Analysis
Airbnbs have become a new trend in the last decade. We all have been on vacation and wanted a place to feel like home without the feeling of being in a hotel. Especially with large groups, people can save money and stay longer in an Airbnb. The chosen Kaggle dataset provides a lot of information including price, ratings, location, and availability.
The goal of this project was to identify the best neighborhood and type of property to purchase in the Boston Area for rental with Airbnb. Using Jupyter Lab, we analyzed the data and created a heat map of all properties by price. We also did some statistical analysis to decide which neighborhood would be the best in which to purchase a property based on price and ratings. We also chose the best property type and number of bedrooms also based on the best pricing and ratings.
Hypothesis & Null Hypothesis
Hypothesis: An Airbnb, two-bedroom apartment located in Jamaica Plain, Boston is a significantly more profitable listing profile to purchase than other neighborhoods based on: Price Overall Rating Score
Null: There is no statistical significance between neighborhoods and property types to purchase in Boston based on price and ratings
Conclusion
Recommended Profile: two-bedroom, condominium in South End, Boston
The team had to first establish that there was a statistical significance between neighborhoods in Boston listed on Airbnb to purchase property in, based on price and ratings. Our analysis found p-values for price (1.85e-119) and ratings (2.19e-30) signalled statistically significant relationships with Boston neighborhoods listed on Airbnb. So the null hypothesis is rejected and we continued our analysis of the data to see the most profitable Boston neighborhood on Airbnb.
Neighborhood
We created box & whisker plots of the median price and ratings for listings in each neighborhood of interest to find which neighborhoods consistently generated high ratings at the highest prices. The analysis found that we rejected our hypothesis as South Boston Waterfront and South End neighborhoods were the only top areas that were top five in both plots; however, South Boston Waterfront has very little listings which we believe skewed the data and decided South End was the more significantly profitable neighborhood.
Property Type
In order to recommend the most profitable property type to purchase in South End, the team created bar graphs of the mean price and ratings for listings in South End to find which property type consistently generated high ratings at the highest prices. The analysis brought the team to reject the hypothesis that an apartment would be the most profitable property type as we found that condominiums perform the best when considering price and ratings. Townhouses had higher ratings but lower prices and Houses had higher prices but lower ratings.
Bedrooms Available
The team conducted analysis based on bar graphs to determine the number of bedrooms available on listings that consistently scored high ratings while listing high prices. Our analysis found that although two and three bedroom listings had comparable mean ratings, three bedrooms listed significantly higher prices. With three-bedrooms listings being such a low percentage (2.6%) of the properties listed in South End, the data team accepted the hypothesis that a two-bedroom property would be significantly more profitable.
Beyond the Data
Further analysis that could have been done was a value analysis in order to determine the most profitable price range to list the properties once purchased. From the current data the team would be able to pull the five listings in South End with the highest prices and begin analysis from there.
Additional information not included in this dataset the team could use is data related to maintenance costs of similar Airbnb listings in the area. Once the properties were listed, property management will also affect profitability so further analysis could’ve been done to determine best host practices.
The current dataset would allow the team to analyze response rate and response time of hosts as well as strictness of cancellation policy against overall rating scores. Additional qualitative data on guests compliments and complaints could also offer valuable analysis but was not available in this current dataset.