airbnb

EDA for London Air bnb data nov18

motivation for the roject

To gain a better insight into Airbnb . Using CrispDM analysis principles to answer a few high level questions (*listed below)

outputs

-------> HERE <------------

Packages used

import pandas as pd
import requests
import seaborn as sns
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error
import nltk
from nltk.corpus import stopwords
from sklearn.tree import DecisionTreeRegressor
import re 
import numpy as np
import matplotlib.pyplot as plt
from io import StringIO
from wordcloud import WordCloud

Data source

http://data.insideairbnb.com/united-kingdom/england/london/2018-11-04/data/listings.csv.gz

Questions asked?

• 1)Can we predict price? and what are the important features?
• 2)Can we predict review scores? and what are the important features?
• 3)What are the common words used in describing properties?

results

• Able to predict price within $8 per person median error - and get a useful list of feature importances below
• Able to predict price within 1.4% median error - and get a useful list of feature importances below
• most common description words

acknowledgments

Airbnb, Udacity, Stackoverflow