International Journal of Scientific Research in Computer Science, Engineering and Information Technology

ISSN : 2456-3307 (www.ijsrcseit.com)

doi : https://doi.org/10.32628/CSEIT2390269

Car Insurance System

Ms. M. S. Sawalkar 1, Ruturaj Kumbhar 2 , Krishna Jamkar 2, Mihir Mandlik 2, Harshawardhan Patil 2
*1 Assistant Professor, Department of Computer Engineering, JSPM Narhe Technical Campus, Pune,
Maharashtra, India
2 B.E. in Computer Engineering, JSPM Narhe Technical Campus, Pune, Maharashtra, India

ARTICLEINFO ABSTRACT

Many automotive insurance providers are looking to improve their service

Article History:
for their customers, businesses are starting to adapt and implement various
Accepted: 10 April 2023 methods of analysing data for performance, as a result giving better service
Published: 30 April 2023 for their customers from a better understanding of their needs. The main
focus of this project therefore is targeted at automotive insurance providers
looking to implement this into their business, the project would also be
Publication Issue beneficial to stakeholders. We propose a system built using MERN tech
Volume 10, Issue 2 Stack for a better customers experience, by suggesting them the most
March-April-2023 appropriate cover in time. The requirement for this system is to perform a
Page Number more efficient up-selling than classic marketing campaigns.
623-628 Keywords : Automotive insurance, cover, stakeholders, MERN.

I. INTRODUCTION statistically validated associations between products

and consumers.
In this report, we propose a system built for a better
customers’ experience, by suggesting them the most These tools have been successfully adopted in many
appropriate cover in time. fields; however, not much has been done for the
insurance industry.
For most famous platforms, such as Amazon and
Netflix, users must choose between hundreds or even We are constructing a system for car insurance, to
thousands of products and tend to lose interest very allow agents to optimize up-selling performances, by
quickly if they cannot make a decision. selecting customers who are most likely to subscribe an
additional cover.
Recommendation systems are then essential to give
customers the best experience. The originality of our system is to be suited for the
insurance context.
Recommender systems are typically employed to
support marketing decisions, as they identify While traditional recommendation systems, designed
for online platforms (e.g., e-commerce, videos), are

Copyright: © the author(s), publisher and licensee Technoscience Academy. This is an open-access article distributed under 623
the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use,
distribution, and reproduction in any medium, provided the original work is properly cited
 M. S. Sawalkar et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., January-February-2023, 9 (2) : 623-628

constructed on huge datasets and aim to suggest the selecting customers who are most likely to subscribe an
next best offer, insurance products have specific additional cover. The originality of our system is to be
properties which imply that we must adopt a different suited for the insurance context.
approach.
While traditional recommendation systems, designed
for online platforms (e.g., e-commerce, videos), are
Purpose: up-selling. Our goal is to support the agents
constructed on huge datasets and aim to suggest the
that are and will continue to be the best advisers for
next best offer, insurance products have specific
customers, due to their experience and their
properties which imply that we must adopt a different
knowledge of their portfolio.
approach.

In short, our tool helps them by automatically selecting

Our system combines user generated data and
from their large portfolios the customers most likely to
integrates data in order to which customer should be
augment their insurance coverage, in order to optimize
recommended which cover.
up-selling campaigns for instance.

II. THEORY Agile Development Model was used as the framework

for the development of the system. However, the
structure of this thesis is based on Software
In order to make good decisions, it is necessary to
Development Life Cycle (SDLC) to provide an
possess ample amount of information. However, there
overview of the project.
are several examples showing that too much
information is as bad as inadequate information; it is
called information overload problem. A typical Software Development Life Cycle consists of
the following stages:
Recommender System has been introduced to solve • Planning and requirement analysis: Based on the
this problem. It is very popular and useful concept in objective of the product, a project plan will be
current digital era. It is an information filtering system
conducted. The product feasibility will be assessed
that suggests products and services most relevant to the
User. according to economic resources and the technical
requirements. The quality assurance requirements for
Recommender System has been used widely for the the product will then be set. Risk assessment is also
products and services, intended for entertainment like done in this stage to ensure the project proceeds with
music, books, and movies, online games, restaurants minimum risks.
and completely based on user ratings.

• Defining requirements: Product requirements are to

Recommender systems are typically employed to
support marketing decisions, as they identify be defined and documented clearly in this stage. This
statistically validated associations between products is done through a Software Requirement Specification
and consumers. (SRS) document.

These tools have been successfully adopted in many • Designing the product architecture: Base on the SRS
fields; however, not much has been done for the
document, a Design Document Specification (DDS)
insurance industry.
document is formed. It contains design proposals for
We are constructing a system for car insurance, to the product architecture including the flow of data
allow agents to optimize up-selling performances, by with external, and third-party modules if such modules

Volume 9, Issue 2, March-April-2023 | http://ijsrcseit.com 624

 M. S. Sawalkar et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., January-February-2023, 9 (2) : 623-628

are used. The document will then be reviewed by These phases are planning, designing, building, testing,
important stakeholders to decide on the best approach. reviewing and launching the product. A typical
iteration can last from one to three weeks.
• Developing the product: This is the stage where the
product is built. The programming code is generated Every iteration involves cross-functional teams
according to the DDS document. The programming working simultaneously in their own areas. Agile uses
language is chosen depending on which type of an adaptive approach where there is no detailed
software is being developed. Coding guidelines are planning.
defined by the developer’s organization.
It only clarifies on what feature need to be developed
• Testing the product: In this stage, the software is next and what task is required to complete that feature.
tested for defects. Product defects are reported, tracked Development teams can adapt to a product with
and fixed. The testing phase is repeated until the dynamic requirements.
product reaches the standards defined in the SRS.

• Deployment and Maintenance: After the product is III.DESIGN APPROACH

carefully tested, it is ready to be deployed. The product
is deployed according to the strategy of the
organization. Maintenance is done after the product is
released when new errors are discovered or when
enhancements are needed.

Software Development Life Cycle, or SDLC, widely

used in the software industry as a framework to define
tasks performed in each step of the development
process. The goal of SDLC is to provide the best product
with the resources given.

Agile Development Model is one of the two most

popular SDLC models, with the other being the
Waterfall Development Model.

In opposition to the Waterfall method, the agile

method focusses on process adaptability and customer
satisfaction by rapid delivery of working software
product.

Figure 3: Dataflow Diagram

The Agile model is an incremental process model with
multiple iterations. Each iteration consists of the
fundamental SDLC phases with minor modifications.

Volume 9, Issue 2, March-April-2023 | http://ijsrcseit.com 625

 M. S. Sawalkar et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., January-February-2023, 9 (2) : 623-628

The model we propose the following approach to build IV. RESULTS AND DISCUSSION
the targeted recommendation system, illustrated by
Figure 3.
This thesis details the development process of building
We built a unique dataset from multiple internal data an application that serves as one of these systems. The
sources. This dataset includes information about application is based on ReactJS.
current customers’ car policies (current cover, vehicle’s
All concepts and technologies used in the project are
characteristics, premium amounts), information about
explained in their corresponding sections.
contacts between customers and the insurance
company (phone calls, mails, etc.) and claims rate based The result of this project was a web-based application
on customers’ history (in particular: claims not covered that can be used for Car Insurance Recommendation.
by their current covering). The system acts as a centralized control panel and
enables the user to manage his/her car’s data.
This allows us to build relevant features based on
The main method of development applied in this
existing variables from the initial data. This step is in
project was Agile methodology as it is suitable for the
general based on knowledge of datasets and on
changing requirements of the project.
intuition supported by experts from specific fields
about what could be the most explanatory features. The Agile is an incremental model, consisting of many
iterations. However, this thesis structure follows the
Next Step we take as an input the customers’ data. Its traditional Software Development Life Cycle stages to
outputs for each customer are estimated by using a provide an overview of the entire development
token that represents whether a customer added a process.
claim in the past or not, one year after the extraction The overall aims of this project were to produce a
date of features. claim a new policy holder would produce an
insurance claim in the next year.
Next Step aims to provide which insurance
This is inthe hopes that it can make automotive
cover/guarantee is most likely to be added, among the
insurance more accessible to more drivers, through a
missing covers of the customers. This step answers the
solution that produces a more accurate cover.
question: which additional insurance cover should we
recommend? To make claim it would have to effectively apply the
feature information such as the customer’s location
For instance, some customers are already covered by or region, as well as their personal and vehicle
recommended cover because they subscribed to an old information.
version of car insurance product, whose guarantees While it may havebeen slightly difficult to directly
were defined differently. Some simple business rules reference real-world features that were anonymized,
downstream of the model take into account these this was possible to overcome by analyzing feature’s
particularities. importance through user generated data.

Last Step is list of insurance covers which can be Using user generated data, we are able to produce a
claimed by the user according his/her needs. solution that is more complex than a simple one, and
provide reliable claim cover.

Volume 9, Issue 2, March-April-2023 | http://ijsrcseit.com 626

 M. S. Sawalkar et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., January-February-2023, 9 (2) : 623-628

Figure 4.3 : Final cover plan details.

Figure 4.1 : Homepage of website.

All concepts and technologies used in the project are

explained in their corresponding sections. This
includes the relevant theoretical background such as
client-server model, Content Management System, and
Document Object Model.

The result of this project was a web-based application

that can be used for Car Insurance Recommendation.

The system acts as a centralized control panel and

enables the user to manage his/her car’s data.

The main method of development applied in this

project was Agile methodology as it is suitable for the
changing requirements of the project.

The Agile is an incremental model, consisting of many

iterations. However, this thesis structure follows the
traditional Software Development Life Cycle stages to
Figure 4.2 : Insurance plan covers provided to provide an overview of the entire development
the user. process.

The overall aims of this project were to produce a

claim a new policy holder would produce an
insurance claim in the next year.

This is in the hopes that it can make automotive

insurance more accessible to more drivers, through a
solution that produces a more accurate cover.

Volume 9, Issue 2, March-April-2023 | http://ijsrcseit.com 627

 M. S. Sawalkar et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., January-February-2023, 9 (2) : 623-628

To make claim it would have to effectively apply the Due to the short time given however, we did not
feature information such as the customer’s location or explore in depth all the aspects of desired project we
region, as well as their personal and vehicle would have liked to.
information.
In future we would like to implement the same using
While it may have been slightly difficult to directly machine learning algorithms and train unsupervisedand
reference real-world features that were anonymized, semi-supervised models, to determine whether or not
this was possible to overcome by analyzing feature’s their results could be used to improve the insurance
importance through user generated data. claim predictions.

Using user generated data, we are able to produce a There are also certain aspects we would go further in
solution that is more complex than a simple one, and depth for researching and selecting, such as the choice
provide reliable claim cover. user generated data.

The addition of more modules for generating more

V. CONCLUSION
data would also be a good way to improve on the
project in future, because they can yield great results
In conclusion, the Car Insurance System (CIS) project on various projects.
was a success. It would also be a good idea to train a live system
The CIS achieved all the predefined functions and (perhaps a batch system) which takes in batches
quality requirements. With each implementation, of new customer information, and then produces
there was considerable effort to enhance claims automatically within a reasonable time.
extensibility.
With a live system the data can thereby increase with
As a result, the product had high maintainability, more customer information and this could potentially
modifiability, and scalability. lead to a greater number of claims if there is more data
Deployment and maintenance will be done in the it can also be used for review of the system.
upcoming future and is outside of the scope of this Lastly due to the fact the data was anonymized we
paper. were unable to make discoveries about the meaning in
While the CIS is satisfied with the final result, there its real-world applications, we were only able to make
were still many aspects that can be optimized. claims with regards to common cover statistics.

In our opinion, there are always potential for If there had been more time we would have liked to
improvements in code structure, functional logic, try and specifically cross-reference feature values to
user experience design and application performance. certain statistics with more research, for example :
Life insurance or bike insurance regards to vehicle
We will continue to do code refactoring to improve
and producing claims.
readability, maintainability and reduce complexity.

VI. FUTURE WORK VII. REFERENCES

From carrying out this project we were intrigued by
the results we had obtained from at various different
stages.

Volume 9, Issue 2, March-April-2023 | http://ijsrcseit.com 628

 M. S. Sawalkar et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., January-February-2023, 9 (2) : 623-628

[1]. Bass, L., Clements, P. & Kazman, R. 2003.

Software Architecture in Practice. 2nd ed. Boston:
Addison Wesley.
[2]. Crockford, D. 2008. JavaScript: The Good Parts.
1st ed. O’Reilly Media: O’Reilly
[3]. Flanagan, D. 2011. JavaScript: The Definitive
Guide. 6th ed. Fonenix: O'Reilly
[4]. Mauthe, A. & Thomas, P. 2005. Professional
Content Management Systems: Handling Digital
Media Assets. 2nd ed. John Wiley
[5]. GitHub. 2017. Usage. Webpack. Accessed 2018.
https://github.com/webpack/docs/wiki/usage
[6]. Lindley, C. 2014. Front-end Driven Applications-
A New Approach to Applications. Telerik.
Accessed 2018.
https://developer.telerik.com/featured/front-end-
driven-applications-new-approach-applications/
[7]. React. No date. Accessed 2018. https://reactjs.org/
[8]. Bobriakov, Igor. (2018). Top 10 Data Science Use
Cases in Insurance. Accessed 12/02/2020, from
https://medium.com/activewizards-machine-
learning- company/top-10-data-science-use-
cases-in-insurance-8cade8a13ee1
[9]. React. No date. Introducing JSX. Accessed 2018.
https://reactjs.org/docs/introducingjsx.html
[10]. React. No date. React-Component. Accessed 2018.
https://reactjs.org/docs/react-component.html
[11]. W3C. 2016. Cascading Style Sheets. Accessed
2018. https://www.w3.org/Style/CSS/
[12]. W3C. 2017. HTML 5.2. Accessed 2018.
https://www.w3.org/TR/html52/

Cite this article as :

M. S. Sawalkar, Ruturaj Kumbhar, Krishna Jamkar,
Mihir Mandlik, Harshawardhan Patil, "Car Insurance
System", International Journal of Scientific Research in
Computer Science, Engineering and Information
Technology (IJSRCSEIT), ISSN : 2456-3307, Volume 9,
Issue 2, pp.623-628, March-April-2023. Available at
doi : https://doi.org/10.32628/CSEIT2390269
Journal URL : https://ijsrcseit.com/CSEIT2390269

Volume 9, Issue 2, March-April-2023 | http://ijsrcseit.com 629