HOUSE PRICE PREDICTION

Submitted in partial fulfillment of the requirements of the degree

BACHELOR OF ENGINEERING IN COMPUTER
ENGINEERING
By

GROUP MEMBERS ROLL NUMBERS

YADAV SATYAM GOVIND 62

GUPTA ARUNKUMAR 17
SURYANATH

MAKHJANKAR ARSALAN 33
KHALIL

BOAT ZAHEER SULEMAN 04

Supervisor
Prof. Bhushan Jadhav

DEPARTMENT OF COMPUTER ENGINEERING

Gharda Institute of Technology
A/P: Lavel, Tal: Khed, Dist: Ratnagiri, 415708
Mumbai University
[2024-25]
 CERTIFICATE

This is to certify that the Mini Project entitled “HOUSE PRICE

PREDICTION” is a bonafide work of Yadav Satyam Govind(62),Gupta

Arunkumar Suryanath(17), Makhajankar Arsalan Khalil(33), Boat

Zaheer Suleman(04) submitted to the University of Mumbai in partial
fulfillment of the requirement for the award of the degree of “Bachelor of
Engineering” in “Computer Engineering”.

Prof. Bhushan Jadhav

Supervisor

Prof. J.V Khalkar Prof. Dr. P. B. Patil

Head of Department Principal
Table of Contents
Abstract.................................................................................................................................................. 1
Acknowledgement ................................................................................................................................. 2
1.Introduction ........................................................................................................................................ 3
1.1 Motivation.................................................................................................................................... 4
1.2 Problem Statement & Objective ................................................................................................ 5
2.Literature Survey ............................................................................................................................... 6
2.1 Limitations of Existing System .................................................................................................. 6
2.2 Survey of Existing System .......................................................................................................... 7
2.3 Mini Project Contribution.......................................................................................................... 8
3.Proposed System .............................................................................................................................. 10
3.1 Introduction ............................................................................................................................... 10
3.2 Architecture/ Framework ........................................................................................................ 11
3.3 Details of Hardware & Software ............................................................................................. 13
3.4 Experiment and Results ........................................................................................................... 14
3.5 Conclusion & Future Scope ..................................................................................................... 21
4.References ......................................................................................................................................... 22
 Abstract

The House Price Prediction project aims to estimate property values using machine learning

techniques, helping buyers, sellers, and investors make informed decisions. By leveraging a

Kaggle dataset, the project analyzes various real estate factors such as location, square footage,

number of rooms, and market trends to predict house prices accurately. The machine learning

models, including Linear Regression, Random Forest, and XGBoost, are developed and trained

using Amazon SageMaker to ensure high accuracy and efficiency.

The project infrastructure is built using AWS services, where S3 is used for data storage, and

the final model outputs are integrated into an interactive website hosted on WordPress via AWS

Lightsail. This website displays insights, data visualizations, and real-time house price

predictions, providing users with a seamless experience. The combination of machine learning

and cloud computing ensures scalability, making the system efficient for real-world

applications.

1
 Acknowledgement

I would like to express my sincere gratitude to everyone who contributed to the successful
completion of this House Price Prediction project.

First and foremost, I extend my heartfelt thanks to my mentors and instructors for their
invaluable guidance, encouragement, and support throughout the project. Their insights into
machine learning, AWS services, and WordPress have been instrumental in shaping this work.

I would also like to acknowledge Amazon Web Services (AWS) for providing powerful cloud
computing tools, including SageMaker, S3, and Lightsail, which enabled the seamless
execution of this project. Additionally, I am grateful to Kaggle for providing the real estate
dataset that served as the foundation for the predictive model.

Lastly, I appreciate my peers, friends, and family for their continuous encouragement and
motivation, which kept me focused and determined to complete this project successfully.

Thank you all for your support and inspiration.

2
 1.Introduction

In today’s real estate market, predicting house prices accurately is crucial for buyers, sellers,
and investors. Traditional valuation methods often rely on subjective opinions or outdated data,
leading to inconsistent pricing. To address this challenge, machine learning offers a data-driven
approach that enhances accuracy and reliability in house price predictions.

This House Price Prediction project leverages machine learning models to estimate property
values based on key features such as location, square footage, number of rooms, and market
trends. The project utilizes a dataset from Kaggle, where various algorithms, including Linear
Regression, Random Forest, and XGBoost, are trained to provide precise predictions.

The implementation is powered by Amazon Web Services (AWS), with SageMaker used for
model development, S3 for data storage, and WordPress on AWS Lightsail for website
deployment. The interactive website presents real-time predictions, visualizations, and
insights, making it accessible for users.

This project demonstrates how machine learning and cloud computing can revolutionize the
real estate industry by offering an efficient, scalable, and automated solution for property
valuation.

3
1.1 Motivation

The real estate market is dynamic, with house prices fluctuating due to various
factors such as location, demand, economic conditions, and property features.
Traditional property valuation methods are often time-consuming, inconsistent,
and prone to human bias. This inspired the need for a data-driven, automated, and
accurate house price prediction system using machine learning.

The motivation behind this project is to leverage artificial intelligence and cloud
computing to simplify real estate decision-making. By developing a predictive
model, buyers can make informed purchasing decisions, sellers can price their
properties competitively, and investors can analyze market trends effectively.

Furthermore, the integration of AWS services ensures that the system is scalable,
secure, and accessible through an interactive WordPress-based website. This
project aims to bridge the gap between technology and real estate, making
property valuation more efficient, accurate, and user-friendly.

4
1.2 Problem Statement & Objective

Problem Statement:

The real estate market is highly dynamic, with house prices influenced by multiple factors such
as location, property size, number of rooms, and market demand. Traditional methods of price
estimation often rely on subjective analysis, leading to inconsistencies and inaccuracies. This
lack of reliable pricing information makes it difficult for buyers, sellers, and investors to make
informed decisions.

To address this issue, there is a need for a data-driven approach that utilizes machine learning
to predict house prices accurately based on historical data and real estate trends.

Objective

The primary objective of this project is to develop a machine learning-based house price
prediction system that provides accurate property valuations. The key goals include:

• Developing predictive models using Linear Regression, Random Forest, and XGBoost
to estimate house prices.

• Utilizing AWS SageMaker for training and deploying machine learning models
efficiently.

• Storing and managing data securely using AWS S3.

• Creating an interactive website using WordPress on AWS Lightsail, where users can
access real-time predictions and insights.

• Enhancing decision-making for buyers, sellers, and investors by providing reliable,

data-driven house price estimates.

This project aims to bridge the gap between technology and real estate, making property
valuation more accurate, scalable, and accessible.

5
 2.Literature Survey

House price prediction has been widely studied using both traditional statistical methods and
modern machine learning techniques. Early approaches relied on statistical models such as
Hedonic Pricing Models and Multiple Linear Regression (MLR), which considered factors like
location, size, and property features. However, these methods struggled with complex,
nonlinear relationships in data, leading to less accurate predictions. With advancements in
artificial intelligence, machine learning algorithms such as Random Forest, Support Vector
Machines (SVM), and XGBoost have significantly improved prediction accuracy by capturing
intricate patterns within datasets. Additionally, the integration of cloud computing has enabled
large-scale data processing and real-time model deployment. AWS SageMaker is commonly
used for training predictive models, while AWS S3 provides efficient data storage. Many real
estate platforms, including Zillow and Redfin, have successfully implemented AI-driven
models to estimate property values. This project builds on these advancements by leveraging
machine learning, AWS services, and a WordPress-based web interface to provide a scalable
and accurate house price prediction system.

2.1 Limitations of Existing System

Despite advancements in house price prediction models, existing systems still face several
limitations:

Limited Accuracy – Traditional statistical models, such as Multiple Linear Regression, fail to
capture complex relationships between real estate factors, leading to less accurate predictions.

Lack of Real-Time Updates – Many existing systems do not incorporate real-time market
trends, making predictions outdated and less relevant.

Data Availability and Quality – House price prediction models depend heavily on high-
quality datasets. Missing or inconsistent data can significantly impact the accuracy of
predictions.Overfitting in Machine Learning Models – Some machine learning models,
particularly neural networks, may overfit on training data, reducing their ability to generalize
well to new property listings.

High Computational Costs – Advanced AI models, especially deep learning techniques,

require substantial computational power, making them expensive to deploy and maintain.

6
Scalability Issues – Many existing house price prediction platforms struggle with scaling their
models to handle large datasets efficiently, especially when deployed on limited
infrastructure.

User Accessibility and Interpretability – Most AI-driven house price estimation tools provide
numerical outputs without explanatory insights, making it difficult for users to understand
how prices are determined.

This project aims to address these challenges by integrating machine learning with cloud-
based solutions using AWS SageMaker, S3, and Lightsail, ensuring accurate, scalable, and
user-friendly house price predictions.

2.2 Survey of Existing System

Several real estate platforms and machine learning models have been developed to predict
house prices. These existing systems utilize various statistical and AI-based techniques to
estimate property values based on factors like location, size, and market trends. Below are some
well-known systems and their methodologies:

1. Zillow Zestimate – One of the most popular real estate pricing models, Zillow
Zestimate uses machine learning and big data analytics to estimate house prices. It
considers features like past sales data, tax history, and property characteristics.
However, its accuracy depends on the availability and quality of local real estate data.

2. Redfin Estimate – This model utilizes neural networks and deep learning to predict
property prices. Unlike traditional models, Redfin’s approach continuously updates
based on new listings and market fluctuations. However, the model is limited to regions
where sufficient real estate data is available.

3. Hedonic Pricing Models – A traditional statistical method that estimates property prices
based on economic and structural features. These models are simple but struggle to
handle nonlinear relationships in data, reducing accuracy.

4. Multiple Listing Services (MLS) – Real estate agencies rely on MLS databases to
determine house values based on recent sales and market trends. While useful, these
services often lack predictive analytics, making them reactive rather than proactive.

5. Machine Learning-Based Prediction Models – Research studies and Kaggle

competitions have explored machine learning algorithms such as Linear Regression,
7
 Random Forest, XGBoost, and Support Vector Machines (SVM) to predict house
prices. These models outperform traditional statistical approaches but require large

2.3 Mini Project Contribution

Team Member 1 – Yadav Satyam Govind

(Role: Project Lead & Machine Learning Engineer)

• Led the overall planning, execution, and coordination of the project.

• Developed and implemented the machine learning model for house price prediction
using Amazon SageMaker.
• Managed dataset preprocessing, feature engineering, and model evaluation.

Team Member 2 – Gupta Arunkumar Suryanath

(Role: Data Analyst & Researcher)

• Collected and analyzed the Kaggle dataset used for house price prediction.
• Performed data cleaning, visualization, and exploratory data analysis (EDA) to extract
meaningful insights.
• Assisted in selecting the most relevant features for the predictive model.

Team Member 3 – Makhjankar Arsalan Khalil

Role: Web Developer & AWS Deployment Specialist

• Designed and developed the WordPress-based interactive website using AWS

Lightsail.

• Integrated S3 storage for image hosting and ensured smooth embedding of visual
outputs.

• Assisted in the deployment of the trained machine learning model on the web platform.

Team Member 4 – Boat Zaheer Suleman

8
(Role: Documentation & Presentation Manager)

• Compiled and structured the project report, documentation, and findings.

• Created AWS setup guides with step-by-step screenshots for better understanding.
• Designed the PowerPoint presentation for project demonstration.

9
 3.Proposed System
The House Price Prediction System is designed to provide accurate and data-driven property
value estimates using machine learning and cloud computing. Unlike traditional pricing
models, the proposed system integrates Amazon SageMaker for model training, AWS S3 for
secure data storage, and WordPress on AWS Lightsail for an interactive user interface. The
system processes real estate datasets by performing data cleaning, feature selection, and
exploratory data analysis (EDA) to ensure high-quality predictions. Advanced machine
learning algorithms such as Linear Regression, Random Forest, and XGBoost are used to
analyze key property attributes and market trends, delivering more precise price estimates.

To enhance usability, the project features a WordPress-powered website where users can input
property details and instantly receive predictions, along with interactive data visualizations to
aid decision-making. The cloud-based infrastructure ensures scalability, real-time data updates,
and efficient handling of large datasets. By leveraging AWS services, the system offers higher
accuracy, improved performance, and secure data management. This innovative approach
bridges the gap between technology and real estate, helping buyers, sellers, and investors make
informed decisions based on reliable, AI-driven insights.

3.1 Introduction

The House Price Prediction System is designed to provide an accurate, automated, and data-
driven approach to estimating property values. Traditional house pricing methods often rely
on manual evaluation, which can be subjective and inconsistent. To overcome these
challenges, this system integrates machine learning algorithms with cloud-based solutions to
analyze real estate data and generate reliable price predictions.

The proposed system leverages Amazon SageMaker for model training, AWS S3 for secure
data storage, and WordPress on AWS Lightsail for an interactive web interface. It utilizes
advanced machine learning models such as Linear Regression, Random Forest, and XGBoost
to process large datasets, identify patterns, and make precise predictions based on key
property attributes.

With a user-friendly website, users can input property details and instantly receive estimated
house prices, making the system accessible to buyers, sellers, and investors. The integration

10
of cloud computing ensures scalability, high performance, and real-time data processing,
making this system an efficient and innovative solution for the real estate industry.

3.2 Architecture/ Framework

The House Price Prediction System follows a structured architecture that integrates machine
learning, cloud computing, and a web-based interface to provide accurate real estate price
predictions. The system is designed to be scalable, efficient, and user-friendly. Below is an
overview of its architectural framework:

1. Data Collection & Storage

• The dataset, sourced from Kaggle, consists of various features such as location, size,
number of rooms, and market trends.

• The dataset is stored securely in Amazon S3, ensuring easy access and scalability.

2. Data Preprocessing & Analysis

• Data Cleaning & Feature Selection: Unnecessary or missing values are handled, and
relevant features are selected.

• Exploratory Data Analysis (EDA): Statistical insights and data visualizations are
generated to understand trends.

3. Machine Learning Model Training (Amazon SageMaker)

• Algorithm Selection: Models such as Linear Regression, Random Forest, and XGBoost
are trained and evaluated.

• Model Optimization: Hyperparameter tuning and performance evaluation are

performed to select the best model.

• Deployment: The trained model is deployed using Amazon SageMaker, ensuring

seamless integration with the website.

4. Web Application Development (AWS Lightsail & WordPress)

• A WordPress-based website is developed and hosted on AWS Lightsail to provide an

interactive user interface.

• The website allows users to input house details and get instant price predictions.

11
 • Data Visualizations from model outputs are embedded to enhance decision-making.

5. User Interaction & Prediction Retrieval

• Users interact with the WordPress website to enter property details.

• The system processes the input and queries the deployed machine learning model.

• The predicted house price is displayed along with relevant insights.

6. Cloud Infrastructure & Scalability

• Amazon S3 is used for data and image storage.

• AWS Lightsail ensures a stable and scalable web hosting environment.

• Amazon SageMaker handles model training and deployment efficiently.

This architecture ensures high accuracy, scalability, and real-time predictions, making it a
robust solution for house price estimation.

12
3.3 Details of Hardware & Software

Hardware Requirements

1. Local Machine (For Development & Testing)

o Processor: Intel Core i5/i7 or AMD equivalent

o RAM: Minimum 8GB (Recommended 16GB for large datasets)

o Storage: Minimum 256GB SSD (Recommended 512GB SSD)

o GPU: Optional, but NVIDIA GPU is preferred for faster model training

2. Cloud Infrastructure (AWS Services Used)

o Amazon SageMaker – For machine learning model training and deployment

o AWS S3 – Cloud storage for datasets and output images

o AWS Lightsail – Hosting the WordPress website for interactive user access

o AWS EC2 (Optional) – For additional computing power if needed

Software Requirements

Development Environment

• Jupyter Notebook (Amazon SageMaker Notebook Instance) – For coding and training
models
• Python (Version 3.x) – Primary programming language
• Libraries Used: Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn, XGBoost

Machine Learning & Data Processing

• Amazon SageMaker – To train, test, and deploy machine learning models

• Scikit-Learn, XGBoost, TensorFlow/PyTorch – For model development
• Matplotlib & Seaborn – For data visualization

Website Development & Deployment

13
 • WordPress (Hosted on AWS Lightsail) – For user interaction and displaying
predictions
• Astra Theme & Elementor Plugin – To enhance UI/UX design
• PHP & MySQL – For backend and database management

Version Control & Collaboration

• Git & GitHub – For source code management

• Google Drive / AWS S3 – For storing datasets and project documentation

3.4 Experiment and Results

Downloading House Price Prediction dataset using Kaggle

We are going to work using train.csv file

14
Using Kaggle, we are going to paste the code of Notebook in Jupiter Lab which is going to be
created using Amazon Sagemaker

Opening Amazon SageMaker and Opening Jupiter Lab

Using conda_python3 kernel

15
Opening Jupiter and importing train.csv

Naming the file name as House_Price_Prediction.ipynb and running the commands

Extracting all the output images from the notebook

16
Saving all the output image in the local device folder

Opening Amazon S3 and creating a bucket named housepriceprebucket

Adding an AWS bucket policy for Public Access

17
Uploading all the files which we have saved in the local device folder including some images
of the mini project members

Using Amazon LightSail, Creating a WordPress named WordPress-1

Opening it by creating login credential (username and password) by putting some commands
in the WordPress Terminal

18
Copying Public IPv4 address and pasting it into new tab

Now login to WordPress using the login credential

Now for our Mini Project we have used Astra as the theme

19
Now importing images in the WordPress from images by Object URL uploaded in the S3
bucket created

Adding the link in the WordPress

After designing the WordPress Website:

Homepage

20
3.5 Conclusion & Future Scope

The House Price Prediction System successfully integrates machine learning, cloud computing,
and web technologies to offer an efficient, data-driven approach to estimating real estate prices.
By leveraging Amazon SageMaker for model training, AWS S3 for data storage, and
WordPress on AWS Lightsail for a user-friendly interface, the system provides accurate price
predictions based on key property attributes such as location, size, and amenities. The cloud-
based infrastructure ensures scalability, security, and accessibility, making it a valuable tool
for homebuyers, sellers, investors, and real estate professionals seeking data-backed decisions.

Looking ahead, the system has the potential for several enhancements. Integrating real-time
market trends from real estate platforms can further improve the accuracy of predictions.
Expanding the dataset to cover multiple cities and regions will make the model more versatile
and applicable to a broader audience. Advanced machine learning techniques, including deep
learning models, can refine price estimations by capturing complex patterns in real estate data.
Additionally, enhancing the user experience with features like price comparisons, investment
recommendations, and mortgage calculators will make the platform more comprehensive.
Future developments could also include a mobile application for better accessibility and
integration with popular real estate portals like Zillow, Realtor, or Redfin to provide users with
live property listings and insights.

With these future improvements, the House Price Prediction System can evolve into a powerful
AI-driven real estate analytics platform, bridging the gap between technology and the real
estate market while enabling smarter, data-driven property decisions.

21
 4.References

• Kaggle. "House Price Prediction Dataset." Available at: https://www.kaggle.com/

• Amazon Web Services (AWS). "Amazon SageMaker Documentation." Available at:
https://docs.aws.amazon.com/sagemaker/
• Amazon Web Services (AWS). "Amazon S3 Documentation." Available at:
https://docs.aws.amazon.com/s3/
• Amazon Web Services (AWS). "AWS Lightsail Documentation." Available at:
https://docs.aws.amazon.com/lightsail/
• Scikit-Learn. "Machine Learning in Python." Available at: https://scikit-learn.org/
• XGBoost. "Extreme Gradient Boosting Library." Available at:
https://xgboost.readthedocs.io/
• WordPress. "Building Websites with WordPress." Available at:
https://wordpress.org/
• NumPy. "Scientific Computing with Python." Available at: https://numpy.org/
• Pandas. "Data Analysis and Manipulation in Python." Available at:
https://pandas.pydata.org/
• Seaborn. "Data Visualization Library in Python." Available at:
https://seaborn.pydata.org/

22