Heart Disease Prediction

Introduction

1.1 Overview: -

The Heart Disease Prediction project leverages machine learning to assess an individual’s risk of
heart disease based on medical and lifestyle factors. It uses a dataset with key attributes like age,
cholesterol levels, blood pressure, and more to train predictive models such as logistic regression,
decision trees, or neural networks. The system provides a user-friendly interface for input and
analysis, helping medical professionals and individuals make informed decisions. This project aims
to enhance early detection, reduce risks, and support healthcare by improving diagnostic accuracy
and preventive care.

Heart disease remains one of the leading causes of mortality worldwide. Early detection and timely
medical intervention are crucial in reducing the risk and severity of heart-related complications. In
this project, we aim to develop a Heart Disease Prediction System using Machine Learning
techniques to assist in the early identification of patients who are at risk.

This project leverages a dataset containing medical parameters such as age, gender, blood pressure,
cholesterol levels, maximum heart rate, chest pain type, and other clinical attributes. By applying
various machine learning algorithms on this data, we build a model that can accurately predict the
presence of heart disease based on input health metrics.

The core objectives of the project are:

 To analyze the correlation between various medical parameters and heart disease.

 To train machine learning models to predict the likelihood of heart disease.

 Toevaluate model accuracy and performance using metrics such as confusion matrix,
accuracy score, and heatmaps.

 Toprovide a user-friendly interface that takes patient data as input and displays prediction
results visually through probability charts and graphs.

Page | 1
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

1.2 Objective of the Project: -

The objective of the Heart Disease Prediction project is to develop a reliable system that utilizes
machine learning algorithms to predict the likelihood of heart disease in individuals. By analysing
medical data such as age, blood pressure, cholesterol levels, and lifestyle factors, the system aims
to provide early detection and risk assessment. This project helps healthcare professionals make
informed decisions, enabling timely intervention and prevention. Additionally, it enhances
awareness among individuals about their heart health, ultimately reducing mortality rates through
proactive measures and improved healthcare management.

1.3 Project Category: -

a) Operating System:
Since Python and machine learning libraries are cross-platform, you can develop and run this
project on any OS.

b) Front-End for Project:

o Tkinter (Built-in Python library for simple desktop UI)
o PyQt or PySide (Advanced GUI with better design and responsiveness)
o Kivy (For cross-platform mobile & desktop applications)

Page | 2
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

System Analysis & Design

2.1 System Analysis and Design

Heart disease is one of the leading causes of death worldwide. Early detection can help in taking
preventive measures and providing timely treatment. This project aims to predict the probability of
heart disease using Machine Learning (ML) by analyzing patient data stored in a CSV file.

1. Feasibility Study

The feasibility study evaluates whether the project is practical, achievable, and beneficial.
This project aims to predict the probability of heart disease using machine learning models
applied to patient data stored in a CSV file. The study considers multiple factors such as cost,
technology, and resources.
Technical Viability: Uses widely available Python libraries like Scikit-Learn, Pandas, and
Flask/Streamlet for visualization.
Operational Viability: Can be implemented in hospitals, clinics, or as a personal health
assessment tool.
Economic Viability: Requires only a computer with Python installed, making it cost-
effective.

2. Economic Feasibility

Economic feasibility ensures that the project is financially viable and does not require
excessive investment.
Hardware Requirements: A basic computer (Windows/Linux/Mac) with at least 8GB RAM
is sufficient.
Software Requirements: Free and open-source tools such as Python, Jupyter Notebook, and
Scikit-learn reduce costs.
Implementation Costs: Minimal, as Python libraries are free. If deployed on a cloud server,
a small hosting fee may apply.
Return on Investment (ROI): High, as it helps in early disease detection, reducing
healthcare costs for patients and hospitals.

Page | 3
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

3. Technical Feasibility

Technical feasibility examines whether the technology used in the project is sufficient for
implementation.
Programming Language: Python (widely supported and has rich ML libraries).
Machine Learning Algorithms:
o Random Forest (for high accuracy)
o Logistic Regression (for probability estimation)
o Neural Networks (for advanced predictions)
Dataset: Uses CSV files with patient medical records.

4. Operational Feasibility

Operational feasibility determines whether the project can function effectively in real-world
scenarios.
User-Friendly Interface: A web-based interface using Flask/Streamlit makes it easy to use.
Accuracy & Performance: Achieves over 80% accuracy with proper ML model tuning.
Integration with Healthcare Systems: Can be connected to hospital databases for real-time
predictions.
Scalability: Can handle large patient datasets and be expanded with cloud deployment for
remote access.

5. Resource Availability

This section ensures that all necessary resources (hardware, software, and human expertise)
are available.
Hardware Requirements:
o Computer with at least 8GB RAM, i5/i7 processor (for faster model training).
o Optional: GPU (if using deep learning models).
Software & Tools:
o Python (Programming language)
o Scikit-Learn, Pandas, NumPy, Matplotlib (For ML and data analysis)
o Flask/Streamlit (For web-based UI)

Page | 4
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

6. Requirement Analysis

Functional Requirements:
o Import and process patient data from a CSV file
o Train machine learning models
o Predict the probability of heart disease
o Display results in a readable format (text/graph)

Non-Functional Requirements:
o Accuracy and reliability of predictions
o Fast processing time
o Easy-to-use interface (CLI or Web UI)
o Secure handling of patient data (if sensitive data is used)

7. Design Modules

Data Input Module: Reads CSV or user input values

Preprocessing Module: Cleans, normalizes, and transforms data
Model Module: Trains and uses machine learning models for prediction
Result Module: Displays heart disease probability
Optional Web UI: Streamlet or Flask front-end for real-time input and output

Page | 5
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

2.2 Project Scheduling and Cost Estimation

Cost Estimation
Software cost comprises a relatively small portion of the overall cost of a computer-based heart
disease prediction system. However, it is a crucial component that needs accurate estimation for
efficient resource allocation and project planning.
A number of factors were taken into account while estimating the cost of this project, including:
o Human Resources (developers, data scientists, testers),
o Technical Requirements (skills, tools, libraries),
o Hardware and Software Availability (laptops, internet, required s/w),
Page | 6
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
o Project Complexity and Model Evaluation Time.

1) Effort Estimation

Effort estimation refers to the total person-hours required for the successful completion of the
Heart Disease Prediction project. This includes activities such as:
o Requirement gathering and analysis
o Dataset collection and preprocessing
o Model selection and training
o Front-end and back-end development
o Integration of ML model with the application
o Testing and debugging
o Preparation of documentation and user manual

2) Hardware Required Estimation

The hardware requirement estimation includes the cost of all physical components used during
the development of the Heart Disease Prediction system. This consists of the following:
Development Machines:
o Personal Computers with minimum configuration:
 Intel i5 Processor
 8 GB RAM
 512 GB SSD
 GPU (optional for model training acceleration)
Internet Connectivity:
o Required for dataset download, library installations, and deployment.

Page | 7
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

2.4 Dataflow Diagram, E-R Diagram:

 DFD Level 0

 DFD Level 1

Page | 8
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

 ER-Diagram

+--------------+ +-------------+
| Patient |◄────────────┐ | Visualization|
+--------------+ │ +-------------+
/ Attributes \ │ ▲
/ ID, Age, BP... \ │ │
/ \ ▼ │
+----------------+ +------------+ "Generated From"
| CSV Database |──────▶| Prediction |◄──────────────────┐
+----------------+ +------------+ │
▲ │
│ "Has" │
▼ │
+---------+ │
| Model |◄─────────────────────┘
+---------+
(Algorithm,
Page | 9
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

Accuracy,
etc.)

❖ Activity
Diagram:

Page | 10
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

System Implementation

3.1 Data Dictionary

Field Data
Description
Name Type
age Integer Age of the patient in years
sex Integer Gender (1 = male; 0 = female)
Chest pain type (0 = typical angina, 1 = atypical angina, 2 =
cp Integer
non-anginal pain, 3 = asymptomatic)
trestbps Integer Resting blood pressure (in mm Hg)
chol Integer Serum cholesterol in mg/dl
fbs Integer Fasting blood sugar > 120 mg/dl (1 = true; 0 = false)
Page | 11
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

Field Data
Description
Name Type
Resting electrocardiographic results (0 = normal, 1 = ST-T
restecg Integer
abnormality, 2 = left ventricular hypertrophy)
thalach Integer Maximum heart rate achieved
exang Integer Exercise-induced angina (1 = yes; 0 = no)
oldpeak Float ST depression induced by exercise relative to rest
Slope of the peak exercise ST segment (0 = upsloping, 1 = flat,
slope Integer
2 = downsloping)
ca Integer Number of major vessels colored by fluoroscopy (0–3)
Thalassemia (1 = normal, 2 = fixed defect, 3 = reversible
thal Integer
defect)
Diagnosis of heart disease (0 = no disease; 1 = presence of
target Integer
disease)

1. System Architecture Overview

The heart disease prediction system is implemented as a machine learning-based web

application that takes user input (medical parameters), processes the data, and predicts the
likelihood of heart disease. The system comprises:
 Frontend: For user input and result display (optional if CLI or Jupyter-based).
 Backend/Model: Python ML model that predicts based on trained data.
 Data Processing Layer: Preprocessing, scaling, and transformation.
 Visualization Module: Bar chart & heatmap to show prediction and correlation.
 Evaluation Layer: Displays model accuracy.

2. Modules in the System

a. Data Input Module
 Users can either:
o Manually input health parameters (like age, cholesterol, blood pressure), or
o Upload a CSV file containing test data.
 Data fields include:
Page | 12
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
o Age, Sex, Chest Pain Type, Resting BP, Cholesterol, etc.

b. Data Preprocessing
 Missing Value Handling: Fill/drop missing or null values.
 Encoding: Convert categorical data to numerical using Label.
 Feature Scaling: Normalize data using StandardScaler or MinMaxScaler.
 Splitting: Dataset is split into training and test sets using train_test_split().

c. Model Training Module

 Machine Learning algorithms used (choose one or multiple for comparison):
o Logistic Regression
o Random Forest Classifier
o K-Nearest Neighbors
o Support Vector Machine (SVM)

d. Prediction Module
 Takes the processed user input.
 Loads the trained model.
 Predicts whether the person is likely to have heart disease.
 Returns result as:
o "Positive" (Likely to have heart disease)
o "Negative" (Unlikely to have heart disease)

e. Visualization Module
 Bar Chart: Displays prediction probability.
 Heatmap: Shows correlation between different medical features (using seaborn or
matplotlib).

f. Model Evaluation Module

 Evaluates model using metrics:

Page | 13
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
o Accuracy, Precision, Recall, F1-score
 Confusion matrix visualized to show TP, FP, TN, FN.

3.2 Modularization Detail

The different modules of this site are as follows:

1. Target Distribution Module

Purpose: Stores and manages patient health records.

 Input: heart_disease_data.csv
 Functionality:
o Load the data using pandas
o Clean and preprocess (handle missing values, normalize, etc.)
o Split into features (X) and target (y)
Sample Fields from CSV:
o Age, Sex, Chest Pain Type, Blood Pressure, Cholesterol, Fasting Blood Sugar, ECG
Page | 14
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
results, Heart Rate, etc.

2. Co-relation Module

Purpose: Uses ML algorithms to predict the probability of heart disease.

Algorithm Options:
o Logistic Regression
o Random Forest
o K-Nearest Neighbors (KNN)

3. Accuracy Module

Purpose: Graphically represent data insights and model results.

 Bar Chart:
o Count of patients with and without heart disease
 Heatmap:
o Correlation matrix of patient features
 Accuracy Chart / Score:
o Display ML model’s performance

3.3 Design

1. Target Distribution Module

Page | 15
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

2. Co-relation Module

Page | 16
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

3. Accuracy Module

Page | 17
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

Roles and responsibility

Page | 18
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

 Project Lead / Coordinator

Responsibilities:
 Oversaw the end-to-end implementation of the Heart Disease Prediction System.
 Coordinated task distribution among team members.
 Managed the dataset lifecycle and ensured adherence to the project timeline.
 Acted as the primary point of communication for event participation and presentation
coordination.

 Data Analyst & Visualization Expert

Responsibilities:
 Explored the dataset using pandas and generated descriptive statistics.
 Used seaborn and matplotlib to create:
 Count plots for target distribution.
 Heatmaps for correlation analysis.
 Leveraged plotly.express to build interactive visualizations for comparing model
accuracy.

 Machine Learning Engineer

Responsibilities:
 Preprocessed data using StandardScaler and handled categorical features with one-hot
encoding (pd.get_dummies).
 Built and evaluated traditional ML models:
 K-Nearest Neighbors (KNN)
 Support Vector Machine (SVM)
 Logistic Regression
 Evaluated each model using accuracy, precision, recall, F1-score, and AUC.

 Testing & Evaluation Specialist

Responsibilities:
Page | 19
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
 Defined and implemented a reusable evaluation function for all models.
 Validated predictions using metrics from sklearn.metrics.
 Assessed model behavior and performance on the validation set to avoid overfitting.

 Documentation & Report Writer

Responsibilities:
 Documented all phases of the project: data preparation, model training,
evaluation, and visualization.
 Compiled findings and results into a structured report.
 Created presentation slides for project demonstrations and event
evaluation.

Testing
Page | 20
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

4. Testing report
4.1 Importance of Testing
Testing is a critical phase in software development, ensuring that the system performs as
expected and is free of bugs. In the Heart Disease Prediction project, testing is vital to:
o Validate the accuracy of predictions made by the machine learning model.
o Ensure data is correctly read and processed from the CSV file.
o Verify that visual components (bar chart, heat map) work correctly.
o Confirm that the application runs without errors and handles edge cases.
o Build confidence in the system’s reliability and accuracy in real-world scenarios.

4.2 The Steps in the Software Testing

The testing in this project followed these essential steps:
1. Requirement Analysis
Understand and define what needs to be tested, including input data formats and
prediction accuracy.
2. Test Planning
Outline the types of tests to be conducted, tools required (e.g., Python libraries),
and performance benchmarks.
3. Test Case Design
Develop test cases to check:
 Correct data loading from CSV
 Valid ML predictions
 Proper rendering of bar charts and heat maps
 Accuracy computation and display
4. Test Environment Setup
Prepare the environment using:
 Python 3.x
 Pandas, Matplotlib, Seaborn, Sklearn
 Jupyter Notebook or any IDE
5. Test Execution
Run the test cases and record outcomes. Make sure plots, results, and data handling
behave as expected.
6. Defect Reporting & Re-testing
Report issues found and correct them. Re-run tests to ensure fixes work without
causing other problems.

4.3 Software Testing Techniques

Some techniques applied during testing include:

Page | 21
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
o Black Box Testing
Testing without knowledge of internal code — used to validate predictions and output
visuals.
o White Box Testing
Testing with understanding of internal logic — used to verify machine learning model logic
and data flow.
o Unit Testing
Individual components (e.g., data loading, model prediction, accuracy function) are tested
separately.
o Integration Testing
Ensures modules (data + model + visualization) work together seamlessly.

4.4 Types of Testing Used

Type of Testing Description

Testing individual Python functions (data loader,
Unit Testing
prediction function)
Ensures CSV data, ML model, and plots work
Integration Testing
together
System Testing Full end-to-end testing of the prediction system
Performance Checks model response time and rendering of
Testing charts
Validates the correctness of heart disease
Accuracy Testing
predictions
Visualization Ensures bar chart and heat map show correct
Testing results

5.3 Test Case

Page | 22
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

Module 1: Data Input (CSV File Load)

Test Case
TC ID Input Expected Output Status
Description
Load CSV file Valid heart.csv Data frame loads with correct
TC_01 Pass
successfully file number of records
Handle file not
TC_02 Invalid file path Display error message Pass
found error
Handle missing/null CSV with Nulls handled or rows
TC_03 Pass
values missing data dropped

Module 2: Data Processing & Model Prediction

TC ID Test Case Description Input Expected Output Status

Encode categorical Columns like All categorical data is
TC_04 Pass
values correctly sex, cp, thal encoded numerically
Split dataset into 70-80% train, 20-30% test
TC_05 Full dataset Pass
train/test sets split
Train the model Model is trained without
TC_06 Training dataset Pass
successfully error
Make prediction for New patient Probability or binary
TC_07 Pass
given patient data input prediction output
Predict for edge case Extreme
TC_08 Valid output (0 or 1) Pass
inputs low/high values

Module 3: Visualization

Page | 23
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

TC ID Test Case Description Input Expected Output Status

Display bar chart for heart Model Bar chart showing count of
TC_09 Pass
disease count output 0 vs 1
Display heat map for Correlation heatmap with
TC_10 Dataframe Pass
correlation matrix color gradient
Accuracy in percentage
TC_11 Display model accuracy Test data Pass
(e.g., 85%)

Module 4: Accuracy and Evaluation

TC ID Test Case Description Input Expected Output Status

Calculate accuracy of Accuracy score (e.g.,
TC_12 Test dataset Pass
trained model 0.85 or 85%)
Display confusion matrix Predictions and Precision, Recall, F1
TC_13 Pass
and classification report actual values Score
Evaluate model using Logistic, Random Comparative accuracy
TC_14 Pass
multiple algorithms Forest, SVM and best model selected

Future Scope

The Heart Disease Prediction system using Python and Machine Learning presents great potential for
Page | 24
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
growth and improvement. Below are some of the key future scopes for enhancing the project:

1. Integration with Real-Time Health Monitoring Systems

The project can be integrated with wearable health monitoring devices to collect real-time data such
as heart rate, blood pressure, and ECG readings. This can allow the model to make live predictions
and alert users immediately in case of risk.

2. Deployment as a Mobile Application

By deploying the model as a web or mobile application, it can be made accessible to hospitals,
clinics, and even individual users. This can increase its usability in real-life scenarios and reach a
broader audience.

3. Incorporation of More Advanced Machine Learning Algorithms

Currently, the model uses basic ML algorithms for prediction. In the future, more advanced models
like XGBoost, Random Forest, or Deep Learning (ANN/CNN) can be implemented for
improved accuracy and prediction power.

4. Larger and More Diverse Dataset

The current model uses data from a CSV file. With access to a larger and more diverse dataset
including different regions, age groups, and health conditions, the model’s generalization and
accuracy can be significantly improved.

Bibliography

Page | 25
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction

Bibliography
1. Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., ... &
Duchesnay, É. (2011).
Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12, 2825–2830.
https://scikit-learn.org/
2. Hunter, J. D. (2007).
Matplotlib: A 2D Graphics Environment. Computing in Science & Engineering, 9(3), 90–95.
https://matplotlib.org/
3. Python Software Foundation.
Python Language Reference, Version 3.9.
https://www.python.org/
4. Various online tutorials and blogs on machine learning and data visualization in Python, including:
o Towards Data Science (https://towardsdatascience.com)
o Kaggle Notebooks (https://www.kaggle.com)

 Conclusion
The Heart Disease Prediction project effectively demonstrates how Python and Machine Learning
Page | 26
Department of Information Technology, NSCOEM, Korti-Pandharpur
 Heart Disease Prediction
algorithms can be utilized to analyse patient health data and predict the probability of heart disease.
The system leverages a CSV dataset containing various patient health measurements such as age,
cholesterol level, blood pressure, chest pain type, and more.
By preprocessing the data and applying a suitable Machine Learning model, the system predicts
whether a patient is likely to have heart disease. A bar chart is used to visualize the number of
patients with and without heart disease, providing a clear graphical insight into the dataset.
Additionally, a heatmap is generated to show the correlation between different medical features,
helping to understand which factors most influence heart disease prediction.
Finally, the system evaluates the performance of the model by calculating and displaying its accuracy
score, which gives an indication of how well the model is performing on unseen data.
Overall, this project highlights the importance and effectiveness of machine learning in the healthcare
domain, offering a supportive tool for early diagnosis and better decision-making in the prevention and
treatment of heart disease.

Page | 27
Department of Information Technology, NSCOEM, Korti-Pandharpur