Abstract
Throughout the years, mobile applications are becoming more popular. Many drivers usually go to car
dealerships, which have a limited variety of cars, and salespeople at the car dealership can use
psychological skills to get a driver to buy a car from the dealership (Dratch, 2018). Essentially, the job
of the car salesmen is to ensure the person who is looking to buy a car purchases a car at the
dealership, because the more sales, the more profit for that particular car dealership, and businesses
are there to maximize profit. This can lead to car buyers buying a car which may not be ideal for them,
leaving the buyer unsatisfied. In this project, I have created an android application which uses
collaborative based filtering and content-based filtering which can be used to figure out what the
user’s ideal car may possibly be. The content-based filtering and collaborative based filtering worked
what I would consider relatively well, although improvements could be made to the content-based
filtering and collaborative based filtering.
,1.1 Problem definition
There are many kinds of cars. Cars can be differentiated by many factors such as manufacturers,
models, year made, colour, speed, number of seats, price and more. With so many factors to take
into consideration, it could be difficult for people to decide which car they would want. Many people
will purchase different kinds of cars depending on these factors. These factors are what makes it
difficult to decide which car to purchase (some of these factors I have outlined in the abstract page
above). ((Armstrong, 2018) Shows a few important factors when it comes to buying cars.)
When customers go to car dealerships of car manufacturer’s in which the user is limited to specific
cars of the car manufacturer, which also does not include all possible lists of cars the manufacturer
manufactures. If a customer goes to a car dealership the salesperson’s job is to the cars which are
available, which kind of seems bias of the salesperson, as they will not recommend cars from other
car manufacturers which could be a better match to the customer. By having a car recommendation
application, it would allow for users who are looking to purchase a car to have a good guideline of
what their possible ideal car could be.
Customers also may want to wait till a car depreciates in value to reach a certain price. There are
many websites which show current prices of cars. Giving an estimate of a car’s price in a few years
could be ideal for users who want a car and willing to wait for the car to reduce in price. New cars
depreciate very quickly (Hull, 2018) after a few years which is why people may want to wait a few
years before purchasing a car. If a person purchases a car many years later after it’s initial release
date (for example buying a car which was released in the year 2005 and buying the vehicle in 2020) ,
it’s likely the person would need to resort to used cars, as it’s unlikely the customer will find an
older car which is brand new.
The average length of time a driver keeps a new vehicle is around 6 years (DeMuro, 2019). When
customers buy a car in person or online user’s previous car information is never used for possible
recommendations, usually, on websites, the user just filters specific manufactures, models, price and
as such of cars to find the car they would like which is time-consuming (Autotrader.co.uk, 2020)).
When customers are recommended cars by websites or car dealerships, usually considered if the
customer has similar features which they consider when buying a car to another customer is not
considered. It could be possible that few customers have some similarities, and by doing so it could
be possible to recommend a car in which one of those customers could be possibly considering
purchasing, and then recommend this car to the other customer.
To consider the problems outlined above, I will need to create a program which needs to be
designed, implemented, tested and deployed. This system will be an android application which can
run on the Android operating system and will also include a database which stores data of different
classes such as user and cars. I will be using the java programming language when creating the
android application.
,In brief, there are some pros to the project. No affiliation with car manufacturers/ dealerships: As we
don’t sell and/or buy cars in the application customers are more likely to trust our car
recommendations Another pro is that consideration of previous cars owned by a person, could
prove useful when recommending news cars, which isn’t usually taken into consideration, especially
on online websites which buy and sell cars. Another pro is when buying a vehicle, a person may want
to wait for a new car model to arrive or waiting specific periods of the month in which there is a
lower demand for buying vehicles which results in big discounts (Tromans, 2018).
There are also a few possible cons to the projects. One con is that when viewing a car in person, you
can physically see the car and tell some of the features which cannot be told using an application, for
example trying to tell a person how comfort a car is subjective and can be difficult to describe, the
best way of knowing how comfortable a car would be simply to sit inside the car. Another con is that
using the information of the previous car to recommend new cars may not be useful, as the car
owner may want a car which is completely different from their previous car, which could result in
incorrect recommendations. Another con is that It is difficult to predict the depreciation rate of an
item, and therefore calculations could be incorrect leading to incorrect results.
, 1.1 AIMS AND OBJECTIVES
The aim of this project is to create an android application which recommends users which car they
should consider purchasing next, this will be done using an algorithm(s) which works effectively to a
certain extent and taking into consideration deterministic factors when it comes to cars. The
objectives are as follows:
1) Create an android application which takes information as input from a user and outputs a
recommended car to the user.
2) Create a database which stores, update and deletes user’s personal data and contains data
of cars
3) Implement an algorithm which recommends a car to a user taking into consideration of
user’s personal data.
4) Research literature of related work, to find possible ways in which the android application
could be created and implemented.
5) Carry out tests on the android application to test if the application works as expected
6) Evaluate the android application against the requirements.
Throughout the years, mobile applications are becoming more popular. Many drivers usually go to car
dealerships, which have a limited variety of cars, and salespeople at the car dealership can use
psychological skills to get a driver to buy a car from the dealership (Dratch, 2018). Essentially, the job
of the car salesmen is to ensure the person who is looking to buy a car purchases a car at the
dealership, because the more sales, the more profit for that particular car dealership, and businesses
are there to maximize profit. This can lead to car buyers buying a car which may not be ideal for them,
leaving the buyer unsatisfied. In this project, I have created an android application which uses
collaborative based filtering and content-based filtering which can be used to figure out what the
user’s ideal car may possibly be. The content-based filtering and collaborative based filtering worked
what I would consider relatively well, although improvements could be made to the content-based
filtering and collaborative based filtering.
,1.1 Problem definition
There are many kinds of cars. Cars can be differentiated by many factors such as manufacturers,
models, year made, colour, speed, number of seats, price and more. With so many factors to take
into consideration, it could be difficult for people to decide which car they would want. Many people
will purchase different kinds of cars depending on these factors. These factors are what makes it
difficult to decide which car to purchase (some of these factors I have outlined in the abstract page
above). ((Armstrong, 2018) Shows a few important factors when it comes to buying cars.)
When customers go to car dealerships of car manufacturer’s in which the user is limited to specific
cars of the car manufacturer, which also does not include all possible lists of cars the manufacturer
manufactures. If a customer goes to a car dealership the salesperson’s job is to the cars which are
available, which kind of seems bias of the salesperson, as they will not recommend cars from other
car manufacturers which could be a better match to the customer. By having a car recommendation
application, it would allow for users who are looking to purchase a car to have a good guideline of
what their possible ideal car could be.
Customers also may want to wait till a car depreciates in value to reach a certain price. There are
many websites which show current prices of cars. Giving an estimate of a car’s price in a few years
could be ideal for users who want a car and willing to wait for the car to reduce in price. New cars
depreciate very quickly (Hull, 2018) after a few years which is why people may want to wait a few
years before purchasing a car. If a person purchases a car many years later after it’s initial release
date (for example buying a car which was released in the year 2005 and buying the vehicle in 2020) ,
it’s likely the person would need to resort to used cars, as it’s unlikely the customer will find an
older car which is brand new.
The average length of time a driver keeps a new vehicle is around 6 years (DeMuro, 2019). When
customers buy a car in person or online user’s previous car information is never used for possible
recommendations, usually, on websites, the user just filters specific manufactures, models, price and
as such of cars to find the car they would like which is time-consuming (Autotrader.co.uk, 2020)).
When customers are recommended cars by websites or car dealerships, usually considered if the
customer has similar features which they consider when buying a car to another customer is not
considered. It could be possible that few customers have some similarities, and by doing so it could
be possible to recommend a car in which one of those customers could be possibly considering
purchasing, and then recommend this car to the other customer.
To consider the problems outlined above, I will need to create a program which needs to be
designed, implemented, tested and deployed. This system will be an android application which can
run on the Android operating system and will also include a database which stores data of different
classes such as user and cars. I will be using the java programming language when creating the
android application.
,In brief, there are some pros to the project. No affiliation with car manufacturers/ dealerships: As we
don’t sell and/or buy cars in the application customers are more likely to trust our car
recommendations Another pro is that consideration of previous cars owned by a person, could
prove useful when recommending news cars, which isn’t usually taken into consideration, especially
on online websites which buy and sell cars. Another pro is when buying a vehicle, a person may want
to wait for a new car model to arrive or waiting specific periods of the month in which there is a
lower demand for buying vehicles which results in big discounts (Tromans, 2018).
There are also a few possible cons to the projects. One con is that when viewing a car in person, you
can physically see the car and tell some of the features which cannot be told using an application, for
example trying to tell a person how comfort a car is subjective and can be difficult to describe, the
best way of knowing how comfortable a car would be simply to sit inside the car. Another con is that
using the information of the previous car to recommend new cars may not be useful, as the car
owner may want a car which is completely different from their previous car, which could result in
incorrect recommendations. Another con is that It is difficult to predict the depreciation rate of an
item, and therefore calculations could be incorrect leading to incorrect results.
, 1.1 AIMS AND OBJECTIVES
The aim of this project is to create an android application which recommends users which car they
should consider purchasing next, this will be done using an algorithm(s) which works effectively to a
certain extent and taking into consideration deterministic factors when it comes to cars. The
objectives are as follows:
1) Create an android application which takes information as input from a user and outputs a
recommended car to the user.
2) Create a database which stores, update and deletes user’s personal data and contains data
of cars
3) Implement an algorithm which recommends a car to a user taking into consideration of
user’s personal data.
4) Research literature of related work, to find possible ways in which the android application
could be created and implemented.
5) Carry out tests on the android application to test if the application works as expected
6) Evaluate the android application against the requirements.
