MIS272 Predictive Analytics

Final Submission due: 12 August 8PM

Description
The purpose of this assignment is to develop your ability to (i) explore patterns in a business dataset utilising
descriptive data mining concepts, and (ii) apply predictive models to address questions relevant to a
particular business context.
The business context for this assignment is the international tourism sector, focusing on providers of tourist
accommodation. Organisations such as AirBnB provide a digital platform that tourists can use to rent
properties in particular locations around the world. The properties are owned by private individuals
(property hosts), and AirBnB takes a commission for bookings via their digital platform.
AirBnB approached you to generate some insights into rental listings in the country of Denmark. AirBnB
provided you with a dataset of 23,941 listings of rentals for the period of Nov 2016-October 2019. The data
set includes the following information:
• Property / room id, its type and price per minimum nights’ stay (in US$)
• Id of the property host (the person owning it or renting it out)
• Property description, geo-location (latitude & longitude) and neighbourhood
• Date first listed, number of reviews recorded so far (since listing)
• Minimum number of nights to be booked (if applicable)
• The number of occupants allowed/can be accommodated
• Overall satisfaction (average from all people who have rented the property)
AirBnB would like you to use RapidMiner to address the following using the provided data set:
Task A: Create a geospatial (map-based) visualisation of all rental properties in Denmark, using their geo-
locations to automatically categorize metropolitan versus regional properties. Metropolitan properties are in
close proximity (refer how this should be calculated below) to any the four major Danish cities (København,
Aalborg, Aarhus, Odense), whereas regional properties are further away. According to this definition, how
many rental properties are metropolitan versus regional?
Task B: AirBnB wants to know if there are important differences between metropolitan and regional
properties in the dataset. Explore this from the perspective of people staying at the rentals (define).
Task C: AirBnB wants to develop pricing guidelines for prospective property hosts. Rentals are classified as
“budget” if the price per night stay is less than $US90. Based on relevant attributes in the dataset, develop
different predictive models to classify a rental property as either budget or not. Evaluate the performance of
the models, indicating the best predictive model. At least one of the models must enable you to describe the
relevant input attribute(s) and their values/ranges that predict whether properties are classified as budget or
not.
The dataset for this assignment is available on CloudDeakin.
Calculating approximate proximity to city centres
The latitude and longitude of the city centres (traditionally, the city hall or “rådhus” in Danish) are as follows:
City centre Latitude (in degrees) Longitude (in degrees)
Aalborg 57.0482743583805 9.920841473627085
Aarhus 56.1526288533459 10.203044717777567
Odense 55.39635401658666 10.389520385212887
København (Copenhagen) 55.67542242571594 12.570195385221885
A rental property is considered in close proximity of a city centre if both of the following conditions apply:
• the difference between the latitude of the rental property and the latitude of the city centre is less
than 0.2 in absolute terms
• the difference between the longitude of the rental property and the longitude of the city centre is
less than 0.4 in absolute terms
For example, a rental property is in close proximity of Aalborg if the following expression is true:
• abs (57.0482743583805-latitude) < 0.2 && abs (9.920841473627085-longitude) < 0.4