Airbnb Data analysis — Boston or Seattle

Logo — Credits Airbnb

Airbnb has collected good amount of data about different properties and their prices.

These data if effectively used, can give us lots of insights about the hosts and buyers and we can predict the future property prices as well.

In this story I will be analyzing the Airbnb data of Boston and Seattle and we will compare various features and also predict the house prices by using simple linear regression model.

We will try to find answers for below three questions

1. Property price comparison between the regions.
2. Trend in price over period of time
3. Various features that affect the price and their correlation.
4. Predicting future prices by building a model using the existing features.

Property prices are not going to be the same at all places. There are many criteria that determine the price of a property.

Lets look how the prices fair between Boston and Seattle:

Seattle visuals

Boston visuals

2. Price Trend:

Below trend chart shows Boston prices decreases a bit over time whereas Seattle property prices were increasing over time.

Overall we could see Boston property average prices are significantly higher when compared to the Seattle.

Price trend over time

3. Property price depends on different features of the property.

From below visual we can clearly see price is highly correlated with the number of different rooms and property size.

4. Predicting Prices:

So this dataset of Boston and Seattle has two years of data and various parameters have been collected. This dataset can be used for predicting new property prices effectively.

I have used linear regression modelling for this prediction and please visit my github repository to see more on that:

Vijaiananth/Airbnb-Boston-Seattle-DataAnalysis

Conclusion:

From above analysis we infer

1. Cleaning fee, monthly price, weekly price and security deposit seems to be higher in Seattle.
2. price seems to be decreasing over time in Boston whereas it increases over time in Seattle.
3. property features and facilities do have bigger impact in its price.
4. by creating a linear regression model we were able to predict the scores with approximately same score between the test and train data set.

Credits: Kaggle, Airbnb and Udacity