Bridge the Gap

Inspiration

The short-term vacation rental market has exploded in the past few years, with large growth expected in the next 6 years. In order to capitalize on emerging markets, we need to analyze the current state of the short-term rental market in Dublin, Ireland and identify key features of what constitutes a successful sale.

What it does

Our project cleans the dataset and wrangles outliers in order to improve data usability. We then perform exploratory data analysis to establish initial findings about the dataset. Next, we dive deeper and use Random Forests/Logistic Regression to identify the features of what makes a good transaction. This analysis allows us to develop actionable insights that can be used for better motivating market action for suppliers.

How we built it

We primarily used python for this project. We used pandas, numpy, and plotly for exploratory data analysis, then transitioned to sklearn for ML models. The final presentation was constructed in canva.

Challenges we ran into

The dataset was very minimal, with extreme outliers and a limited supply of useable data. Having more weeks/months of data would allow for more accurate predictive forecasting.

Accomplishments that we're proud of

We created machine learning models based on different versions of the data (unprocessed contacts.tsv and processed). Logistic regression and Random Forest models were used achieving accuracy scores of 81-90% at predicting whether or not an individual would book based on their individual factors.

What we learned

Through this analysis, we learned that guests want to bring more people for longer stays and they want to plan longer in advance. Also, hosts need to accommodate guest inquiries as much as possible and Airbnb should have more properties for special times and in Dublin City Centre.

What's next for the project

The dataset however, could be made much better as there are many crucial variables missing. Firstly, a longer time span than two weeks would allow for better models and predictions overall. Additional data on price, room types and neighborhoods of booked properties would be useful. As for what's next, the methods used for Dublin could also be applied to other cities, by accounting for region-specific holidays and popular events.