New York City Airbnb Open Data
In this project, I will explore the NYC Airbnb listings and aim to find out listing that match with upcoming trip
In [0]:
#Import packages and libraries
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
!pip install geopandas
import geopandas as gpd
In [0]:
#Load data
from google.colab import drive
drive.mount ('/content/gdrive')
In [0]:
data = pd.read_csv('gdrive/My Drive/dataset/AB_NYC_2019.csv')
First Inspection of Data
In [0]:
data.head(5)
Out[0]:
In [0]:
data.dtypes, data.columns
Out[0]:
In [0]:
#Change name of 'neighbourhood_group' to 'boroname'
data.rename(columns={'neighbourhood_group':'boroname'}, inplace=True)
In [0]:
#Find out the listings by borough name
plt.figure(figsize=(10,10))
sns.scatterplot(x='longitude', y='latitude', hue='boroname', s=30, data=data)
Out[0]:
Listing in different Boroughs in NYC
In [0]:
# Count listing by borough
borough_count = data.groupby('boroname').agg('count').reset_index()
borough_count
Out[0]:
In [0]:
#Plot the count by borough
fig, ax1 = plt.subplots(1,1, figsize=(6,6))
sns.barplot(x='boroname', y='id', data=borough_count, ax=ax1, palette="plasma" )
ax1.set_title('Number of Listings by Borough', fontsize=15)
ax1.set_xlabel('Borough', fontsize=12)
ax1.set_ylabel('Count', fontsize=12)
ax1.tick_params(axis='both', labelsize=10)
Manhattan have more listings that other regions
In [0]:
#Using geopandas to bring in a base layer of NYC boroughs
nyc = gpd.read_file(gpd.datasets.get_path('nybb'))
nyc.head(5)
Out[0]:
In [0]:
#Rename the 'BoroName' to 'boroname', this help to join data for the next step
nyc.rename(columns={'BoroName':'boroname'}, inplace=True)
bc_geo = nyc.merge(borough_count,on='boroname')
bc_geo
Out[0]:
In [0]:
#Plot into the map count listing by borough
fig,ax = plt.subplots(1,1, figsize=(10,10))
bc_geo.plot(column='id', cmap='viridis_r', alpha=.5, ax=ax, legend=True)
bc_geo.apply(lambda x:ax.annotate(s=x.boroname, color='black', xy=x.geometry.centroid.coords[0],ha='center'), axis=1)
plt.title('Number of Airbnb Listings by NYC Borough')
plt.axis('off')
Out[0]:
We could see that most of listings locate in Brooklyn and Manhattan. In contract, Airbnb is not so popular in Bronx and Staten Island
In [0]:
#import shapely
from shapely import wkt
Since we don't have geometries of neighborhoods, so we could use CSV file from NYC Open Data Site. It has wkt in a geometry column
In [0]:
crs = {'init':'epsg:4326'}
geometry = gpd.points_from_xy(data.longitude, data.latitude)
geo_data = gpd.GeoDataFrame(data,crs=crs,geometry=geometry)
In [0]:
fig,ax = plt.subplots(figsize=(15,15))
nyc.plot(ax=ax,alpha=0.4,edgecolor='black')
geo_data.plot(column='id',ax=ax,legend=True,cmap='plasma',markersize=4)
plt.title("Number of Airbnb Listings")
plt.axis('off')
Out[0]:
For the upcoming trip, I would like to find 'Entire home/apt' option and prefer if it locate in Brooklyn. So let's plot a map to see locations of room type
In [0]:
fig,ax = plt.subplots(figsize=(15,15))
nyc.plot(ax=ax, alpha=0.4, edgecolor='black')
geo_data.plot(column='room_type',ax=ax,legend=True,cmap='plasma',markersize=4)
plt.title('Locations of room type')
plt.axis('off')
Out[0]:
Also find out different room type by borough
In [0]:
plt.figure(figsize=(10,10))
ax = sns.countplot(data['room_type'], hue=data['boroname'], palette='plasma')
Manhattan offer more entire home/apt than other regions. However, I am a student love travelling on budget. Therefore, I will focus on listings in Manhattan which have price under 65
In [0]:
data_manha_65=data[data.boroname=='Manhattan']
data_manha_65.head()
Out[0]:
In [0]:
data_manha_65 = data_manha[data_manha.price <65]
data_manha_65['label']=data_manha_65.apply(lambda x: (x['name'],'price:'+str(x['price'])),axis=1)
data_manha_65.head()
Out[0]:
According to below map, you could not only see location of listings but also comments.
In [0]:
#install folium to create a map
!pip install git+https://github.com/python-visualization/folium
In [0]:
import folium
from folium import plugins
Long=-73.92
Lat=40.86
manha_map=folium.Map([Lat,Long],zoom_start=12)
manha_rooms_map=plugins.MarkerCluster().add_to(manha_map)
for lat,lon,label in zip(data_manha_65.latitude,data_manha_65.longitude,data_manha_65.label):
folium.Marker(location=[lat,lon],icon=None,popup=label).add_to(manha_rooms_map)
manha_map.add_child(manha_rooms_map)
manha_map = folium.Map([Lat,Long],zoom_start=12)
manha_map
Out[0]:
In [0]: