Understand Airbnb rental landscape in Seattle — Data Analysis

For all prospective Airbnb hosts in Seattle, I will answer these questions in this article:

• when to rent to maximise revenue?
• when is the off-peak season for maintenance?
• common group size of Seattle travellers, is it 2 or family or 4 or larger?
• bedroom configurations to maximise booking rates?
• how to achieve a good rating?
• do hosts with higher rating have higher revenue?
• amenities to include?

Get and prepare data

In this article, I will perform exploratory data analysis on the Airbnb dataset gotten from Inside Airbnb.

Fetch Listings data

Our data will be loaded in pandas, comma-separated values (CSV) files can be easily loaded into DataFrame with the read_csv function.
Let us look at what the first 10 rows looks like with pd_listings.head(10):

	id	name	neighbourhood_group_cleansed	latitude	longitude	property_type	room_type	accommodates	bathrooms	bedrooms	beds	amenities	price	guests_included	minimum_nights	number_of_reviews	review_scores_rating	review_scores_accuracy	review_scores_cleanliness	review_scores_checkin	review_scores_communication	review_scores_location	review_scores_value
	241032	Stylish Queen Anne Apartment	Queen Anne	47.63628904	-122.3710252	Apartment	Entire home/apt	4	1	1	1	{TV,"Cable TV",Internet,"Wireless Internet","Air Conditioning",Kitchen,Heating,"Family/Kid Friendly",Washer,Dryer}	$85.00	2	1	207	95	10	10	10	10	9	10
	953595	Bright & Airy Queen Anne Apartment	Queen Anne	47.63912312	-122.3656665	Apartment	Entire home/apt	4	1	1	1	{TV,Internet,"Wireless Internet",Kitchen,"Free Parking on Premises","Buzzer/Wireless Intercom",Heating,"Family/Kid Friendly",Washer,Dryer,"Smoke Detector","Carbon Monoxide Detector","First Aid Kit","Safety Card","Fire Extinguisher",Essentials}	$150.00	1	2	43	96	10	10	10	10	10	10
	3308979	New Modern House-Amazing water view	Queen Anne	47.62972413	-122.3694832	House	Entire home/apt	11	4.5	5	7	{TV,"Cable TV",Internet,"Wireless Internet","Air Conditioning",Kitchen,"Free Parking on Premises","Pets Allowed","Pets live on this property",Dog(s),Cat(s),"Hot Tub","Indoor Fireplace",Heating,"Family/Kid Friendly",Washer,Dryer,"Smoke Detector","Carbon Monoxide Detector",Essentials,Shampoo}	$975.00	10	4	20	97	10	10	10	10	10	10
	7421966	Queen Anne Chateau	Queen Anne	47.6384732	-122.3692791	Apartment	Entire home/apt	3	1	0	2	{Internet,"Wireless Internet",Kitchen,"Indoor Fireplace",Heating,"Family/Kid Friendly",Washer,Dryer,"Smoke Detector","Carbon Monoxide Detector","Safety Card","Fire Extinguisher",Essentials,Shampoo}	$100.00	1	1	0
	278830	Charming craftsman 3 bdm house	Queen Anne	47.6329184	-122.3724706	House	Entire home/apt	6	2	3	3	{TV,"Cable TV",Internet,"Wireless Internet",Kitchen,Heating,"Family/Kid Friendly","Smoke Detector","Carbon Monoxide Detector","First Aid Kit","Fire Extinguisher",Essentials,Shampoo}	$450.00	6	1	38	92	9	9	10	10	9	9
	5956968	Private unit in a 1920s mansion	Queen Anne	47.63052548	-122.3661741	House	Private room	2	1	1	1	{"Wireless Internet","Free Parking on Premises",Heating,"Smoke Detector",Essentials,Shampoo}	$120.00	1	1	17	95	10	10	10	10	10	10
	1909058	Queen Anne Private Bed and Bath	Queen Anne	47.63660545	-122.3685191	House	Private room	2	1	1	1	{"Wireless Internet","Free Parking on Premises",Heating,"Smoke Detector","First Aid Kit",Essentials,Shampoo}	$80.00	1	3	58	99	10	10	10	10	10	10
	856550	Tiny Garden cabin on Queen Anne	Queen Anne	47.64016054	-122.375856	Cabin	Private room	2	1	1	1	{"Wireless Internet","Pets live on this property",Dog(s),Heating,"Family/Kid Friendly",Essentials,Shampoo}	$60.00	1	2	173	97	10	10	10	10	9	10
	4948745	Urban Charm || Downtown || Views	Queen Anne	47.6324103	-122.3572161	Apartment	Private room	2	1	1	1	{TV,"Cable TV",Internet,"Wireless Internet",Kitchen,Breakfast,"Indoor Fireplace",Heating,Washer,Dryer,"Smoke Detector","First Aid Kit","Fire Extinguisher",Essentials,Shampoo}	$90.00	1	3	8	97	10	9	10	9	10	10

view raw seattle_airbnb_open_data-pd_listings.head.csv hosted with ❤ by GitHub

And examine the summary of the numerical data with pd_listings.describe():

		id	latitude	longitude	accommodates	bathrooms	bedrooms	beds	guests_included	minimum_nights	number_of_reviews	review_scores_rating	review_scores_accuracy	review_scores_cleanliness	review_scores_checkin	review_scores_communication	review_scores_location	review_scores_value
	count	3.818000e+03	3818.000000	3818.000000	3818.000000	3802.000000	3812.000000	3817.000000	3818.000000	3818.000000	3818.000000	3171.000000	3160.000000	3165.000000	3160.000000	3167.000000	3163.000000	3162.000000
	mean	5.550111e+06	47.628961	-122.333103	3.349398	1.259469	1.307712	1.735394	1.672603	2.369303	22.223415	94.539262	9.636392	9.556398	9.786709	9.809599	9.608916	9.452245
	std	2.962660e+06	0.043052	0.031745	1.977599	0.590369	0.883395	1.139480	1.311040	16.305902	37.730892	6.606083	0.698031	0.797274	0.595499	0.568211	0.629053	0.750259
	min	3.335000e+03	47.505088	-122.417219	1.000000	0.000000	0.000000	1.000000	0.000000	1.000000	0.000000	20.000000	2.000000	3.000000	2.000000	2.000000	4.000000	2.000000
	25%	3.258256e+06	47.609418	-122.354320	2.000000	1.000000	1.000000	1.000000	1.000000	1.000000	2.000000	93.000000	9.000000	9.000000	10.000000	10.000000	9.000000	9.000000
	50%	6.118244e+06	47.623601	-122.328874	3.000000	1.000000	1.000000	1.000000	1.000000	2.000000	9.000000	96.000000	10.000000	10.000000	10.000000	10.000000	10.000000	10.000000
	75%	8.035127e+06	47.662694	-122.310800	4.000000	1.000000	2.000000	2.000000	2.000000	2.000000	26.000000	99.000000	10.000000	10.000000	10.000000	10.000000	10.000000	10.000000
	max	1.034016e+07	47.733358	-122.240607	16.000000	8.000000	7.000000	15.000000	15.000000	1000.000000	474.000000	100.000000	10.000000	10.000000	10.000000	10.000000	10.000000	10.000000

view raw seattle_airbnb_open_data-pd_listings.describe.csv hosted with ❤ by GitHub

Observations:

• there are 3813 listings in this dataset
• values in the price column contain the dollar symbol ($)
• there are missing values in columns bathrooms, bedrooms, and beds there are missing values in reviews rating columns (review_scores_rating, review_scores_accuracy, review_scores_cleanliness, review_scores_checkin, review_scores_communication, review_scores_location, - review_scores_value)

Data preparation

The column price, which is the price of the listing, it contains the dollar sign ($). We still can’t use it for analysis as it is not a numerical value, so we remove the dollar symbol and convert the values as numeric values:

pd_listings['price'] = pd_listings['price'].str.replace("[$, ]", "").astype("float")

Then replace those empty values with zero:

pd_listings.at[pd_listings['bathrooms'].isnull(), 'bathrooms'] = 0
pd_listings.at[pd_listings['bedrooms'].isnull(), 'bedrooms'] = 0
pd_listings.at[pd_listings['beds'].isnull(), 'beds'] = 0
pd_listings.at[pd_listings['review_scores_rating'].isnull(), 'review_scores_rating'] = 0
pd_listings.at[pd_listings['review_scores_accuracy'].isnull(), 'review_scores_accuracy'] = 0
pd_listings.at[pd_listings['review_scores_cleanliness'].isnull(), 'review_scores_cleanliness'] = 0
pd_listings.at[pd_listings['review_scores_checkin'].isnull(), 'review_scores_checkin'] = 0
pd_listings.at[pd_listings['review_scores_communication'].isnull(), 'review_scores_communication'] = 0
pd_listings.at[pd_listings['review_scores_location'].isnull(), 'review_scores_location'] = 0
pd_listings.at[pd_listings['review_scores_value'].isnull(), 'review_scores_value'] = 0

Lastly, to rename id to listing_id:

pd_listings.rename(columns={'id':'listing_id'}, inplace=True)

Fetch Reviews data

Let us load another CSV file which contains the reviews for each listing. The DataFrame contains the following columns:

• id — identification number for review
• listing_id — identification number for listing which we can join with the above DataFrame
• date — date of the review

Calculate estimated revenue for each listing

I suppose that each review is a successful booking and guests stayed some number of nights. Unfortunately, we do not know the exact number of nights each guest stayed, but we could use the listing’s minimum_nights, to assume each guest stayed at least that minimum number of nights. For each review, price * minimum_nights to get each booking’s revenue:

pd_bookings = pd.merge(pd_reviews, pd_listings, on='listing_id')
pd_bookings['estimated_revenue'] = pd_bookings['price'] * pd_bookings['minimum_nights']

Sum up the revenue of every booking for each listing as estimated revenue per listing:

pd_listings_revenue = pd_bookings[['listing_id','estimated_revenue']].groupby(['listing_id']).sum()

And merged the estimated revenue into the existing DataFrame (listing):

pd_listings = pd.merge(pd_listings, pd_listings_revenue, on='listing_id', how='left')
pd_listings.at[pd_listings['estimated_revenue'].isnull(), 'estimated_revenue'] = 0

And we have our DataFrame ready for some analysis. Each row represents one listing, its attributes, and its estimated revenue:

		listing_id	name	neighbourhood_group_cleansed	latitude	longitude	property_type	room_type	accommodates	bathrooms	bedrooms	beds	amenities	price	guests_included	minimum_nights	number_of_reviews	review_scores_rating	review_scores_accuracy	review_scores_cleanliness	review_scores_checkin	review_scores_communication	review_scores_location	review_scores_value	estimated_revenue
	0	241032	Stylish Queen Anne Apartment	Queen Anne	47.63628904	-122.3710252	Apartment	Entire home/apt	4	1	1	1	{TV,"Cable TV",Internet,"Wireless Internet","Air Conditioning",Kitchen,Heating,"Family/Kid Friendly",Washer,Dryer}	85	2	1	207	95	10	10	10	10	9	10	17595
	1	953595	Bright & Airy Queen Anne Apartment	Queen Anne	47.63912312	-122.3656665	Apartment	Entire home/apt	4	1	1	1	{TV,Internet,"Wireless Internet",Kitchen,"Free Parking on Premises","Buzzer/Wireless Intercom",Heating,"Family/Kid Friendly",Washer,Dryer,"Smoke Detector","Carbon Monoxide Detector","First Aid Kit","Safety Card","Fire Extinguisher",Essentials}	150	1	2	43	96	10	10	10	10	10	10	12900
	2	3308979	New Modern House-Amazing water view	Queen Anne	47.62972413	-122.3694832	House	Entire home/apt	11	4.5	5	7	{TV,"Cable TV",Internet,"Wireless Internet","Air Conditioning",Kitchen,"Free Parking on Premises","Pets Allowed","Pets live on this property",Dog(s),Cat(s),"Hot Tub","Indoor Fireplace",Heating,"Family/Kid Friendly",Washer,Dryer,"Smoke Detector","Carbon Monoxide Detector",Essentials,Shampoo}	975	10	4	20	97	10	10	10	10	10	10	78000
	3	7421966	Queen Anne Chateau	Queen Anne	47.6384732	-122.3692791	Apartment	Entire home/apt	3	1	0	2	{Internet,"Wireless Internet",Kitchen,"Indoor Fireplace",Heating,"Family/Kid Friendly",Washer,Dryer,"Smoke Detector","Carbon Monoxide Detector","Safety Card","Fire Extinguisher",Essentials,Shampoo}	100	1	1	0	0	0	0	0	0	0	0	0
	4	278830	Charming craftsman 3 bdm house	Queen Anne	47.6329184	-122.3724706	House	Entire home/apt	6	2	3	3	{TV,"Cable TV",Internet,"Wireless Internet",Kitchen,Heating,"Family/Kid Friendly","Smoke Detector","Carbon Monoxide Detector","First Aid Kit","Fire Extinguisher",Essentials,Shampoo}	450	6	1	38	92	9	9	10	10	9	9	17100
	5	5956968	Private unit in a 1920s mansion	Queen Anne	47.63052548	-122.3661741	House	Private room	2	1	1	1	{"Wireless Internet","Free Parking on Premises",Heating,"Smoke Detector",Essentials,Shampoo}	120	1	1	17	95	10	10	10	10	10	10	2040
	6	1909058	Queen Anne Private Bed and Bath	Queen Anne	47.63660545	-122.3685191	House	Private room	2	1	1	1	{"Wireless Internet","Free Parking on Premises",Heating,"Smoke Detector","First Aid Kit",Essentials,Shampoo}	80	1	3	58	99	10	10	10	10	10	10	13920
	7	856550	Tiny Garden cabin on Queen Anne	Queen Anne	47.64016054	-122.375856	Cabin	Private room	2	1	1	1	{"Wireless Internet","Pets live on this property",Dog(s),Heating,"Family/Kid Friendly",Essentials,Shampoo}	60	1	2	173	97	10	10	10	10	9	10	20760
	8	4948745	Urban Charm || Downtown || Views	Queen Anne	47.6324103	-122.3572161	Apartment	Private room	2	1	1	1	{TV,"Cable TV",Internet,"Wireless Internet",Kitchen,Breakfast,"Indoor Fireplace",Heating,Washer,Dryer,"Smoke Detector","First Aid Kit","Fire Extinguisher",Essentials,Shampoo}	90	1	3	8	97	10	9	10	9	10	10	2160
	9	2493658	Airy + Bright Queen Anne Apartment	Queen Anne	47.63749172	-122.3668886	Apartment	Entire home/apt	4	1	1	1	{TV,Internet,"Wireless Internet",Kitchen,"Free Parking on Premises","Buzzer/Wireless Intercom",Heating,"Family/Kid Friendly",Washer,Dryer,"Smoke Detector","Carbon Monoxide Detector","First Aid Kit","Safety Card","Fire Extinguisher",Essentials}	150	1	2	32	97	10	10	10	10	10	9	9600

view raw seattle_airbnb_open_data-pd_listings.head-2.csv hosted with ❤ by GitHub

Begin analysis

Revenue by neighbourhoods

This table shows the average revenue of listings in each neighbourhood:

	neighbourhood_group_cleansed	estimated_revenue
	Downtown	7247.666037735849
	Capitol Hill	7064.079365079365
	Beacon Hill	6606.983050847458
	Ballard	6078.726086956522
	Queen Anne	6038.752542372881
	Central Area	4636.371273712737
	Other neighborhoods	4409.843828715365
	Cascade	4075.1348314606744
	Seward Park	4063.5
	Rainier Valley	3827.3459119496856
	Delridge	3641.1898734177216
	Magnolia	3587.8196721311474
	West Seattle	3370.783251231527
	Northgate	2962.3625
	Lake City	2476.4328358208954
	Interbay	2105.7272727272725
	University District	1558.5573770491803

view raw seattle_airbnb_open_data-toplisting-neighbourhood.csv hosted with ❤ by GitHub

Airbnb properties in Downtown, Capitol Hill and Beacon Hill can fetch the highest revenue. It’s shopping and CBD district.

Downtown, Capitol Hill and Beacon Hill can fetch the highest revenue

Popular period of the year to rent?

It would be useful to know the most popular time of the year to rent in Seattle, so Airbnb hosts are able to decide when to rent and when is the time for maintenance.

July, August and September are the best periods to maximise revenue. Months before May are the best time for maintenance work. From October to December is a good time to take a break and enjoy the holidays if they want to.

July, August and September are the best periods to maximise revenue.

Highest revenue listings

These are the top 5 listings with the highest estimated revenue:

		listing_id	number_of_reviews	minimum_nights	accommodates	bedrooms	beds	estimated_revenued
	2617	3594885	8	1000	4	1.0	1.0	1200000.0
	2107	5056580	100	31	2	1.0	1.0	306900.0
	1500	4009508	38	20	5	2.0	2.0	266000.0
	1537	1954452	71	14	2	1.0	1.0	218680.0
	1519	3971934	48	20	3	1.0	1.0	171840.0

view raw seattle_airbnb_open_data-pd_listings-toplistings.csv hosted with ❤ by GitHub

Wow! Looks like our top earners are hosts have minimum nights of 1000. But it might be data anomaly because 1000 nights are kind of extreme, so let’s look at the proportion of listings with different minimum_nights.

	minimum_nights	'# of listings'
	1	1610
	2	1423
	3	480
	4	122
	5	67
	6	16
	7	50
	8	1
	10	5
	11	1
	12	1
	13	2
	14	16
	15	1
	20	11
	21	1
	26	1
	28	1
	29	1
	30	6
	31	1
	1000	1

view raw seattle_airbnb_open_data-minnights-countlistings.csv hosted with ❤ by GitHub

Most hosts have minimum nights of up to a month, the host with 1000 nights, gotta filter it away.

These are the top hosts (up to 7 minimum nights) with the highest estimated revenue.

		listing_id	number_of_reviews	minimum_nights	accommodates	bedrooms	beds	estimated_revenued
	1070	3385421	31	6	4	2.0	2.0	103602.0
	2	3308979	20	4	11	5.0	7.0	78000.0
	3667	53803	41	7	4	3.0	3.0	71750.0
	1592	9460	240	3	2	1.0	1.0	71280.0
	3165	3040278	156	2	4	2.0	2.0	67704.0

view raw seattle_airbnb_open_data-pd_listings-toplistings2.csv hosted with ❤ by GitHub

These are the top hosts (up to 4 minimum nights) with the highest estimated revenue.

		listing_id	number_of_reviews	minimum_nights	accommodates	bedrooms	beds	estimated_revenued
	2	3308979	20	4	11	5.0	7.0	78000.0
	1592	9460	240	3	2	1.0	1.0	71280.0
	3165	3040278	156	2	4	2.0	2.0	67704.0
	3216	481220	164	2	4	1.0	3.0	63960.0
	2922	699596	136	3	2	1.0	1.0	61200.0

view raw seattle_airbnb_open_data-pd_listings-toplistings-le4days.csv hosted with ❤ by GitHub

From these 2 tables, longer minimum nights results in higher revenue. Let's look at the correlation between minimum nights and estimated revenue.

		minimum_nights	estimated_revenue
	minimum_nights	1.0	0.8720842868140248
	estimated_revenue	0.8720842868140248	1.0

view raw seattle_airbnb_open_data-corr-minnightsrevenue.csv hosted with ❤ by GitHub

And the correlation between minimum nights and estimated revenue after removing the listing with 1000 minimum nights.

		minimum_nights	estimated_revenue
	minimum_nights	1.000000	0.199189
	estimated_revenue	0.199189	1.000000

view raw seattle_airbnb_open_data-corr-minnightsrevenue-remove1000nights.csv hosted with ❤ by GitHub

Host with 1000 minimum nights has caused a bais towards higher minimum nights resulted in higher revenue, with a correlation of 87% between minimum nights and revenue. But after removing that host, minimum nights and estimated revenue are not highly correlated, a correlation of 20% between minimum nights and revenue.

Minimum nights and estimated revenue are not highly correlated

Supply and demand — bedroom configurations

As an Airbnb host, it will be good to know if my property is oversaturated in the market. Find out the ratio between the number of listings (supply) to the number of bookings (demand) of different bedroom configurations:

	# of bedrooms	# of listings	# of bookings	ratio
	1.0	2417	61399	25.402978899462145
	0.0	378	8403	22.23015873015873
	2.0	640	10593	16.5515625
	6.0	6	82	13.666666666666666
	3.0	283	3593	12.696113074204947
	5.0	24	205	8.541666666666666
	4.0	69	569	8.246376811594203
	7.0	1	5	5.0

view raw seattle_airbnb_open_data-bedroomcount-supplydemand.csv hosted with ❤ by GitHub

Listings with less than 2 bedrooms are well sought after.

But wait! Properties with no bedrooms, what are kind of properties are these?

	room_type	# of listings
	Entire home/apt	378

view raw seattle_airbnb_open_data-nobedroomtype.csv hosted with ❤ by GitHub

And the number of beds in these properties?

	# of beds	# of listings
	1.0	292
	2.0	76
	3.0	8
	4.0	2

view raw seattle_airbnb_open_data-nobedroombedcount.csv hosted with ❤ by GitHub

All of these properties which no bedrooms are renting the entire apartment, and they do provide at least one bed. Phew~

Supply and demand — guest group configuration

As an Airbnb host, I would also like to know the common group size of Seattle visitors. So as to find out if my property configuration is oversaturated in the market.

	accommodates	# of listings	# of bookings	ratio
	14	3	83	27.666666666666668
	2	1627	42821	26.31899200983405
	3	398	10170	25.55276381909548
	10	25	520	20.8
	4	785	16041	20.4343949044586
	7	52	956	18.384615384615383
	5	184	3221	17.505434782608695
	6	332	5580	16.80722891566265
	12	15	229	15.266666666666667
	1	256	3542	13.8359375
	8	119	1501	12.61344537815126
	16	4	48	12.0
	15	2	19	9.5
	9	13	98	7.538461538461538
	11	3	20	6.666666666666667

view raw seattle_airbnb_open_data-groupsize-supplydemand.csv hosted with ❤ by GitHub

A place which accommodates 14 ranked first (highest supply/demand ratio), but the number of bookings is low (only 83 bookings) as compared to places for 2 or 3 people.

Renting a place for 2 or 3 people will give the host pretty good regular rentals.

Supply and demand — bedroom configurations for 2 to 3

So let us focus on renting properties for 2 to 3 people since more than half travel in a group of this size. Do these guests prefer 1 bedroom or 2 separate bedrooms?

Airbnb bedroom configurations for 2 people:

	bedrooms	# of listings	# of bookings	ratio
	1.0	1613	40897	25.354618722876626
	0.0	261	5376	20.597701149425287
	2.0	9	90	10.0

view raw seattle_airbnb_open_data-bedroomfor2-supplydemand.csv hosted with ❤ by GitHub

Airbnb bedroom configurations for 3people:

	bedrooms	# of listings	# of bookings	ratio
	1.0	311	8214	26.411575562700964
	0.0	53	1396	26.339622641509433
	2.0	34	560	16.470588235294116

view raw seattle_airbnb_open_data-bedroomfor3-supplydemand.csv hosted with ❤ by GitHub

The majority prefers 1 bedroom.

The majority prefers 1 bedroom, less than 1% prefers 2 bedrooms. So for groups of 2s or 3s, they prefer 1 bedroom. But this could be due to the current supply of 2 bedroom properties are low.

What factors matters?

Having good ratings is important for Airbnb hosts. Let us compare how different factors affect overall ratings:

Good communication affects the overall rating and check-in rating

Communication has the highest correlation with the overall rating. Host in Seattle (maybe elsewhere too) needs to be responsive and friendly because good communication tends to get a high overall rating. Good communication also directly impacts the check-in rating.

Does having a good overall rating means the listing will bring in good wealth?

Having a good overall rating has a very small positive correlation with estimated revenue. And having a good rating has almost no impact on the price set by the host.

But still, having a good overall rating is highly recommended.

Amenities

These are the number of Airbnbs in Seattle that provides these amenities:

	amenities	# of listings provide
	Wireless Internet	3667.0
	Heating	3627.0
	Kitchen	3423.0
	Smoke Detector	3281.0
	Essentials	3237.0
	Dryer	2997.0
	Washer	2992.0
	Internet	2811.0

view raw seattle_airbnb_open_data-amenities.csv hosted with ❤ by GitHub

Internet, heating and kitchen are necessities in Seattle.

Smoke detector? I just learnt that the Washington State Building Code has required smoke detectors in all dwellings since 1973.

Summary

So, here is the summary of this article:

Notebook

Check out the codes used in this article!

Loading

Sorry, something went wrong. Reload?

Sorry, we cannot display this file.

Sorry, this file is invalid so it cannot be displayed.

view raw seattle-airbnb-open-data.ipynb hosted with ❤ by GitHub