Inspiration

We were inspired by our own experience as airbnb tenants. To see the requirements and best factors for a well reviewed and profitable airbnb allowed us to look at the business model from the other side of the screen.

What it does

The model analyses and predicts the most important factors for a well-informed business decision about placement, and pricing for an airbnb letting. It creates spatial binning for the airbnbs from some of the most famous cities from around the world and looked at other spatial factors like ammenities in the neighbourhood.

How we built it

The big part of building the model was the extracting relevant data that the model could be trained on. Similar to a landlord needing to understand the pros and cons of their listing, the model focused on extracting the importance of features and predicting how they affect the final price.

Challenges we ran into

The biggest challenge was the poor quality of the data. Starting with multiple datasets having incorrect data i.e. data that not match the scheme, to strongly biased reviews of costumers which blocked us from getting a true review of the quality gotten for price and location.

Accomplishments that we're proud of

The working model with a reasonable confidence which is a huge achievement given relatively small number of data points we had after cleaning. Visualisation was another achievement, with allowing to produce heat-maps of the cities based off of multiple factors and boundaries.

What we learned

The two main hard skills we honed was working with spatial data and presenting them in an accessible manner. As well as analysis of the importance of the features of the model. The big lesson however, was the need for understanding of the limitations of the data before starting the project.

What's next for AirBnB-Sadness

Continually improving the skills we picked up today and probably an afternoon of sleep.

Built With

Share this project:

Updates