Booking prices analysis

Basic info:

  • goal: Verify if the given hotel price limit for a particular city is accurate to the market prices
  • author: Slawomir Drzymala
  • code: github/sdrzymala
  • last update date: 2019-10-27


To achieve the given goal the following steps has been made:

  1. Get hotel/apartments/etc (property) details and prices for given city for the same length of stay for different days
  2. Specify the office location and calculate the distance between the office and each property
  3. Analye the data and check the distribution of price per person per night
    • in total
    • excluding outliers (most luxurious properties)
    • for hotels only
    • for trusted properties only (with review score > 5)
    • for properties that are within the walking distance to the office
    • all together

1.1 Import libraries and configure jupyter

1.2 Get data from other notebooks

1.3 Configure google maps

1.4 Configure other parameters

1.5 Prepare other functions

2.1 Introduction to analysis

Booking searches analysis

On 2019-10-28 the website was searched to create the analysis of prices in the city of Warsaw ( Poland ) for given criteries:

  • Number of nights: 3
  • Number of adults: 1
  • Number of rooms: 1
  • Is business trip: True

The given searches has been done:

  • from 2019-10-28 to 2019-10-31 which is 0 number of days in advance.
  • from 2019-11-04 to 2019-11-07 which is 7 number of days in advance.
  • from 2019-11-11 to 2019-11-14 which is 14 number of days in advance.
  • from 2019-11-18 to 2019-11-21 which is 21 number of days in advance.
  • from 2019-11-25 to 2019-11-28 which is 28 number of days in advance.

As the results the price and not only informations of 2172 properties in total has been collected

To get a better insight the following destination has been defined: Office location in Warsaw, location = (20.9782, 52.2226)

Please note that the following parameters has been setup:

  • Price per person per night limit: 175
  • Walking distance limit (meters): 2500

Please also note that the prices below are shown in PLN

3.1 Overview of the data - number of properties

3.2 Distribution of ppp per night (all)

3.3 Distribution of ppp per night (without outliers)

3.4 Distribution of ppp per night (without outliers) zoom in

3.5 Distribution of ppp per night (without outliers) per search date

3.6 Distribution of ppp per night (without outliers) include ratings

3.7.1 Create function to show points on the map

3.7.2 Create map with all properties

3.7.3 Show map with all properties

Wait a bit till the map will be loaded. If the map won't pop up double click on the empty area and wait again... (this might take some time)

Hoover over the point to see the details about the property including estimated walk duration and the distance
Please note that the blue circle is for reference only and due to the projection issue it might not fit the distance well. This will be only the projection issue and the distance is calculated using the google maps api and should be very acurate.

Blue dot - destination (office) location
Red dots = properties that have the average price per person per night above given limit
Green dots = properties that have the average price per person per night below given limit

3.8 Distribution of ppp per night (without outliers) exclude low ratings and distant properties

4. Outcome and findings

Booking searches analysis

The current limit for Warsaw is 175

Taking into the consideration most recent data and the given conditions:

  • Ignoring the outliers - properties that are much more expensive than the others (based on ZScore)
  • Ignoring properties that are not in the defined walking distance (2500 meters) to the office location
  • Ignoring properties with review score lower than 5
  • Ignoring all the other types of the property apart from Hotels

Out of 2172 there is only 14 properties that match all of the criteries and the price statistics are as follow:
  • Min: ~178
  • Mean: ~307
  • Median: ~315
  • Standard deviation: ~78
  • Max: ~462

Consequently the current limit ( 175 ) is 57% lower than the mean ( 307 ) and 55% lower than the median ( 315 ) comparing to the current prices of the hotels in the neighbours of the office (taking into the consideration above additional conditions). Furthermore there is 0 properties that are matching the given criterias if we would include the given price limit.

For better understanding please take a look at the charts presented above or contact us

5. Export analysis