Inside Airbnb: Listing concentration by host

economy

python

Who is really behind the hosted accounts on Airbnb? A “pretty” idea that was born as a result of high rental prices in San Francisco and ended up taxing its revenue in Ireland.

Published

Nov 3, 2025

Keywords

liberalization

Summary

Airbnb, a platform that offers intermediary services between tenants and property owners, has become a challenge in many cities, as well as a controversial topic for defending liberalization and low regulation of the housing market.

This chart is a replica of the one published by Anna Gordun on towardsdatascience, with the addition of direct data extraction using BeautifulSoup and an alternative to her code using pandas and matplotlib.

Code

# Libraries
# ==========================================
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
import requests
from bs4 import BeautifulSoup
from io import StringIO

# Data Extraction (insideairbnb)
# ==========================================
# Extract url pages in a dataframe
url = "https://insideairbnb.com/get-the-data/"
response = requests.get(url)
response.encoding = 'utf-8'
soup = BeautifulSoup(response.text, 'html.parser')

# Get unique URLs that contain 'listings.csv'
listings_urls = set(a['href'] for a in soup.find_all('a', href=True) if a['href'].endswith('listings.csv'))

# Convert the set back to a DataFrame
df_urls = pd.DataFrame(list(listings_urls), columns=["url"])

# Add city, province and country
df_urls['city'] = df_urls['url'].apply(lambda x: x.strip('/').split('/')[-4].title() if len(x.strip('/').split('/')) >= 4 else None)
df_urls['province'] = df_urls['url'].apply(lambda x: x.strip('/').split('/')[-5].title() if len(x.strip('/').split('/')) >= 5 else None)
df_urls['country'] = df_urls['url'].apply(lambda x: x.strip('/').split('/')[-6].title() if len(x.strip('/').split('/')) >= 6 else None)

# Solve empty values 
df_urls['country'] = df_urls.apply(lambda row: row['city'] if row['province'] == "Data.Insideairbnb.Com" else row['country'], axis=1)
df_urls['country'] = df_urls.apply(lambda row: row['province'] if row['country'] == "Data.Insideairbnb.Com" else row['country'], axis=1)
df_urls['province'] = df_urls.apply(lambda row: row['city'] if row['province'] == "Data.Insideairbnb.Com" else row['province'], axis=1)

# Filter cities and convert to list
urls = df_urls.to_dict(orient='records')

# Create empty list
dfs = []

# Iterate over each url and city in the list of dictionaries
for entry in urls:
    url = entry['url']
    city = entry['city']
    province = entry['province']
    country = entry['country']
    response = requests.get(url)
    df = pd.read_csv(StringIO(response.text), encoding='utf-8')
    df['city'] = city
    df['province'] = province
    df['country'] = country
    dfs.append(df)

# Concatenate dataframes
df = pd.concat(dfs, axis=0, ignore_index=True)

# Data Manipulation
# ==========================================
# Group and category data
host_count = df.groupby(['host_id']).size().reset_index(name='count')
bins = [0, 1, 2, 5, 10, 100, float('inf')]
labels = ['1', '2', '+2', '+5', '+10', '+100']
host_count['host_category'] = pd.cut(host_count['count'], bins=bins, labels=labels, right=True)

# Join and select columns
df = pd.merge(df, host_count, on='host_id', how='left')
df = df[['country', 'id', 'host_id', 'host_name', 'host_category', 'count', 'license']]

# Check if have license
df['license_category'] = df['license'].apply(lambda x: 0 if pd.isna(x) or x == 'Exempt' else 1)

# Grouping by host category
df = df.groupby(['country', 'host_category']).agg(
    {'id': 'count', 'license_category': 'sum'}).reset_index()

# Rename columns
df = df.rename(columns={'id': 'property', 'license_category': 'license'})

# Calculate percents
df['side'] = np.where(df['host_category'].isin(['1', '2']), 'left', 'right')
df['property_percent'] = df['property'] / df.groupby('country')['property'].transform('sum') * 100
df['property_percent'] *= df['side'].eq('left').map({True: -1, False: 1})

# Pivot columns
df_pivot = df.pivot(index="country", columns="host_category", values="property_percent").fillna(0)

# Define column with values for individuals and professionals
df_pivot['total_left'] = df_pivot['1'] + df_pivot['2']
df_pivot['total_right'] = df_pivot['+2'] + df_pivot['+5'] + df_pivot['+10'] + df_pivot['+100']
df_pivot = df_pivot.sort_values(by='total_left', ascending=True)

# Select and order columns
order = ["2", "1", "+2", "+5", "+10", "+100"]
dfplot = df_pivot[order]

print(dfplot)

# Data Visualization
# ==========================================
# Font and style
plt.rcParams.update({'font.family': 'sans-serif', 'font.sans-serif': ['Franklin Gothic'], 'font.size': 9})
sns.set(style="white", palette="muted")

# Palette color
palette = ["#537c78", "#7ba591", "#cc222b", "#f15b4c", "#faa41b", "#ffd45b"]

# Create horizontal stack bar plot
ax = dfplot.plot(kind="barh", stacked=True, figsize=(10, 7), width=0.7, color=palette)

# Add title and labels
ax.text(0, 1.12, f'Who is your Airbnb host?', fontsize=16, fontweight='bold', ha='left', transform=ax.transAxes)
ax.text(0, 1.07, f'% of airbnb listings in European countries managed by singles/multiproperty hosts', fontsize=11, color='#262626', ha='left', transform=ax.transAxes)
ax.set_xlim(-100, 100)
xticks = np.linspace(-75, 75, 7)
plt.xticks(xticks, labels=[f"{abs(int(i))}%" for i in xticks], fontsize=7)
plt.gca().set_ylabel('')
plt.yticks(fontsize=8, color='#282828', fontweight='bold')
plt.grid(axis='x', linestyle='--', color='gray', linewidth=0.5, alpha=0.3)
plt.axvline(x=0, color='#282828', linestyle='-', linewidth=1.5, alpha=0.7)

# Add individual and professional text
plt.text(0.25, 1.02, 'Individual', fontsize=9.5, fontweight='bold', va='center', ha='center', transform=ax.transAxes, color="#537c78")
plt.text(0.75, 1.02, 'Professional', fontsize=9.5, fontweight='bold', va='center', ha='center', transform=ax.transAxes, color="#faa41b")

# Add strict regulation zone
ynum = 14
ax.axvspan(-100, 100, ymin=0, ymax=ynum/len(dfplot), color='gray', alpha=0.1)
plt.axhline(y=ynum-0.5, color='#282828', linestyle='--', linewidth=0.5, alpha=0.3)
plt.text(-100, ynum-0.3, 'Countries with strict Airbnb Regulations', fontsize=6, fontweight='bold', color="gray")

# Add values for individual bars
for i, (city, total) in enumerate(zip(dfplot.index, df_pivot['total_left'])):
    ax.text(total - 1, i, f'{abs(total):.0f}', va='center', ha='right', fontsize=7, color='grey')

# Add values for professional bars
for i, (city, total) in enumerate(zip(dfplot.index, df_pivot['total_right'])):
    ax.text(total + 1, i, f'{total:.0f} ', va='center', ha='left', fontsize=7, color='grey')

# Legend configuration
plt.plot([], [], label="Number of properties per host: ", color='white')
plt.legend(
    loc='lower center', 
    bbox_to_anchor=(0.5, -0.1),
    ncol=7,
    fontsize=8,
    frameon=False,
    handlelength=1,
    handleheight=1,
    borderpad=0.2,
    columnspacing=0.2
)

# Add Data Source
plt.text(0, -0.12, 'Data Source:', 
    transform=plt.gca().transAxes, 
    fontsize=8,
    fontweight='bold',
    color='gray')
space = " " * 23
plt.text(0, -0.12, space + 'Inside Airbnb', 
    transform=plt.gca().transAxes, 
    fontsize=8,
    color='gray')

# Remove spines
for spine in plt.gca().spines.values():
    spine.set_visible(False)
    
# Plot it! :)
plt.show()

Code

# Libraries
# ==========================================
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
import requests
from bs4 import BeautifulSoup
import re
from io import StringIO

# Data Extraction (insideairbnb)
# ==========================================
# Extract url pages in a dataframe
url = "https://insideairbnb.com/get-the-data/"
response = requests.get(url)
response.encoding = 'utf-8'
soup = BeautifulSoup(response.text, 'html.parser')

# Get unique URLs that contain 'listings.csv'
listings_urls = set(a['href'] for a in soup.find_all('a', href=True) if a['href'].endswith('listings.csv'))

# Convert the set back to a DataFrame
df_urls = pd.DataFrame(list(listings_urls), columns=["url"])

# Add city, province and country
df_urls['city'] = df_urls['url'].apply(lambda x: x.strip('/').split('/')[-4].title() if len(x.strip('/').split('/')) >= 4 else None)
df_urls['province'] = df_urls['url'].apply(lambda x: x.strip('/').split('/')[-5].title() if len(x.strip('/').split('/')) >= 5 else None)
df_urls['country'] = df_urls['url'].apply(lambda x: x.strip('/').split('/')[-6].title() if len(x.strip('/').split('/')) >= 6 else None)

# Solve empty values 
df_urls['country'] = df_urls.apply(lambda row: row['city'] if row['province'] == "Data.Insideairbnb.Com" else row['country'], axis=1)
df_urls['country'] = df_urls.apply(lambda row: row['province'] if row['country'] == "Data.Insideairbnb.Com" else row['country'], axis=1)
df_urls['province'] = df_urls.apply(lambda row: row['city'] if row['province'] == "Data.Insideairbnb.Com" else row['province'], axis=1)

# Filter cities and convert to list
cities = {"Girona", "Barcelona", "Amsterdam", "Prague", "Mallorca", "Lisbon", "Malaga", "Athens", "Viena", "Crete", "Florence", "London", "Edinburgh", "Rome", "Paris", "Copenhagen", "Andalucia"}
df_urls = df_urls[df_urls['city'].isin(cities)]
urls = df_urls.to_dict(orient='records')

# Create empty list
dfs = []

# Iterate over each url and city in the list of dictionaries
for entry in urls:
    url = entry['url']
    city = entry['city']
    province = entry['province']
    country = entry['country']
    response = requests.get(url)
    df = pd.read_csv(StringIO(response.text), encoding='utf-8')
    df['city'] = city
    df['province'] = province
    df['country'] = country
    dfs.append(df)

# Concatenate dataframes
df = pd.concat(dfs, axis=0, ignore_index=True)

# Data Manipulation
# ==========================================
# Group and category data
host_count = df.groupby(['host_id']).size().reset_index(name='count')
bins = [0, 1, 2, 5, 10, 100, float('inf')]
labels = ['1', '2', '+2', '+5', '+10', '+100']
host_count['host_category'] = pd.cut(host_count['count'], bins=bins, labels=labels, right=True)

# Join and select columns
df = pd.merge(df, host_count, on='host_id', how='left')
df = df[['city', 'id', 'host_id', 'host_name', 'host_category', 'count', 'license']]

# Check if have license
df['license_category'] = df['license'].apply(lambda x: 0 if pd.isna(x) or x == 'Exempt' else 1)

# Grouping by host category
df = df.groupby(['city', 'host_category']).agg(
    {'id': 'count', 'license_category': 'sum'}).reset_index()

# Rename columns
df = df.rename(columns={'id': 'property', 'license_category': 'license'})

# Calculate percents
df['side'] = np.where(df['host_category'].isin(['1', '2']), 'left', 'right')
df['property_percent'] = df['property'] / df.groupby('city')['property'].transform('sum') * 100
df['property_percent'] *= df['side'].eq('left').map({True: -1, False: 1})

# Pivot columns
df_pivot = df.pivot(index="city", columns="host_category", values="property_percent").fillna(0)

# Define column with values for individuals and professionals
df_pivot['total_left'] = df_pivot['1'] + df_pivot['2']
df_pivot['total_right'] = df_pivot['+2'] + df_pivot['+5'] + df_pivot['+10'] + df_pivot['+100']
df_pivot = df_pivot.sort_values(by='total_left', ascending=True)

# Select and order columns
order = ["2", "1", "+2", "+5", "+10", "+100"]
dfplot = df_pivot[order]

print(dfplot)

# Data Visualization
# ==========================================
# Font and style
plt.rcParams.update({'font.family': 'sans-serif', 'font.sans-serif': ['Franklin Gothic'], 'font.size': 9})
sns.set(style="white", palette="muted")

# Palette color
palette = ["#537c78", "#7ba591", "#cc222b", "#f15b4c", "#faa41b", "#ffd45b"]

# Create horizontal stack bar plot
ax = dfplot.plot(kind="barh", stacked=True, figsize=(10, 6), width=0.7, color=palette)

# Add title and labels
ax.text(0, 1.12, f'Who is your Airbnb host?', fontsize=16, fontweight='bold', ha='left', transform=ax.transAxes)
ax.text(0, 1.07, f'% of airbnb listings in European cities owned by singles/multiproperty hosts', fontsize=11, color='#262626', ha='left', transform=ax.transAxes)
ax.set_xlim(-100, 100)
xticks = np.linspace(-75, 75, 7)
plt.xticks(xticks, labels=[f"{abs(int(i))}%" for i in xticks], fontsize=8)
plt.gca().set_ylabel('')
plt.yticks(fontsize=9, color='#282828', fontweight='bold')
plt.grid(axis='x', linestyle='--', color='gray', linewidth=0.5, alpha=0.3)
plt.axvline(x=0, color='#282828', linestyle='-', linewidth=1.5, alpha=0.7)

# Add individual and professional text
plt.text(0.25, 1.02, 'Individual', fontsize=9.5, fontweight='bold', va='center', ha='center', transform=ax.transAxes, color="#537c78")
plt.text(0.75, 1.02, 'Professional', fontsize=9.5, fontweight='bold', va='center', ha='center', transform=ax.transAxes, color="#faa41b")

# Add strict regulation zone
ynum = 6
ax.axvspan(-100, 100, ymin=0, ymax=ynum/len(dfplot), color='gray', alpha=0.1)
plt.axhline(y=ynum-0.5, color='#282828', linestyle='--', linewidth=0.5, alpha=0.3)
plt.text(-100, ynum-0.4, 'Cities with strict Airbnb Regulations', fontsize=7, fontweight='bold', color="gray")

# Add values for individual bars
for i, (city, total) in enumerate(zip(dfplot.index, df_pivot['total_left'])):
    ax.text(total - 1, i, f'{abs(total):.0f}', va='center', ha='right', fontsize=7, color='grey')

# Add values for professional bars
for i, (city, total) in enumerate(zip(dfplot.index, df_pivot['total_right'])):
    ax.text(total + 1, i, f'{total:.0f} ', va='center', ha='left', fontsize=7, color='grey')

# Legend configuration
plt.plot([], [], label="Number of properties per host: ", color='white')
plt.legend(
    loc='lower center', 
    bbox_to_anchor=(0.5, -0.12),
    ncol=7,
    fontsize=8,
    frameon=False,
    handlelength=1,
    handleheight=1,
    borderpad=0.2,
    columnspacing=0.2
)

# Add Data Source
plt.text(0, -0.135, 'Data Source:', 
    transform=plt.gca().transAxes, 
    fontsize=8,
    fontweight='bold',
    color='gray')
space = " " * 23
plt.text(0, -0.135, space + 'Inside Airbnb', 
    transform=plt.gca().transAxes, 
    fontsize=8,
    color='gray')

# Remove spines
for spine in plt.gca().spines.values():
    spine.set_visible(False)

# Plot it! :)
plt.show()