By: Gabriel Flomo (DS7), Bhavani Rajan (DS8), & Mikio Harman (DS8)
A prediction model that helps you determine the best price to pay based on key features. Find our model here.
The data used was scraped on November 7, 2018 and contained detailed listing data, review data, and calendar data of current lisintgs in Berlin
The data used for the model can be found here.
The data that was cleaned and filtered can be found here
Find our Flask API here
Homepage: This is the landing page for the pred-airbnb web app.
Using JSON: https://pred-airbnb.herokuapp.com/json
{
"cleaning_fee": 30.0,
"accomodates": 3,
"minimum_nights": 4,
"bedrooms": 1,
"bathrooms": 1,
"neighborhood": 1,
"room_type": 15,
"extra_people": 28,
"Laptop_friendly_workspace": 1,
"tv": 1,
"wifi": 1,
"family_kid_friendly": 1,
"smoking_allowed": 0
}
{
"optimal_price": 74.08
}
Find our schema here
Our mean baseline for our y (target) was $67.14
- This helps us gauge how well our model is doing by looking at individual predictions and comparing the actual, mean baseline, and the predicted to find out how well our model did on a case by case basis
Our model had a MAE of 19.32
- This tells us that overall our model will predict a price that is on average $19.32 around the actual price.
A look at the cost of each AirBnB in Berlin
Find visuals like these and many others that help explain the data in meaningful ways here