HeartAI is a fictional organization that operates in different American hospitals under government supervision. Its primary objective is to offer real-time analysis, clinical decision support, and robust artificial intelligence to enhance healthcare. One of the problems faced by HeartAI is the inability of medical staff to predict which patients are at risk of heart attacks among the 2000 patients they attend to.

A solution is to develop a system that accurately and quickly predicts patients with a high probability of experiencing heart strokes, this would greatly benefit the general public and HeartAI. Patients can adjust their diets or seek medical attention promptly with this information. The awareness created by this project would boost HeartAI’s community service and reduce the number of heart attack casualties, which would make it successful for years to come.

Application Benefits

The data product’s development would significantly benefit the general population, reducing mortality rates from heart attacks by diagnosing and providing assistance to distant people at high risk of heart disease. The project's cost is reasonable, and it will help HeartAI and its partners to reach a broader audience.

Application Description

This project is a console-based application that uses machine learning to analyze heart stroke probabilities. We will provide the Jupyter Notebook used to train the software. Additionally, it will be coded in Python, making use of Imbalanced learn, Sklearn, Pandas, and Matplotlib for data visualization purposes.

Data Description

The data used for this model comes from an online source and contains information about different individuals’ cholesterol levels. This information will help identify heart stroke probabilities. After the product's training, HeartAI plans to obtain additional data from populations that will be used to generate predictions on heart disorders.

Objective and Hypotheses

The goal of the data product is to provide a low-cost technique for measuring heart attack rates using cholesterol in restricted populations worldwide. A successful product would enable early treatment and continued management of heart strokes in populations that lack means to identify and adjust their diet.

Our hypothesis is that a model can be taught to predict heart attack probabilities quickly, accurately, and affordably. Furthermore, a successful model will help in understanding and investigating diabetes prevalence in rural areas worldwide.

Methodology

Our project will follow the incremental technique for development. An adaptive methodology will be used instead of a predictive approach to allow for fast development while easily making incremental modifications in the future. This will allow HeartAI to take advantage of the product quickly and respond to in-field staff's feature modification suggestions.

Stakeholder Impact

Developing this data product would significantly impact stakeholders positively. The general population would benefit from reduced heart attack casualties and improved healthcare, something that HeartAI would use to sustain its community service for years to come.

Precautions for Data Usage

The information utilized to prepare the model is accessible to the public and has been modified to ensure the exclusion of private data. However, the information collected in the field may hold information that could lead to identification of individuals. To prevent the collection of data that may reveal identifying personal details, every subject's data will be assigned nameless numbers. Moreover, the collected data will be broad enough to prevent any specific identification.

Technical Proposal (Section B)

Problem Statement

The purpose of this proposal is to introduce a new console application to the remote HeatAI field staff. This application will provide them with the ability to analyze and compare acquired dataset crucial to predicting the likelihood of an individual having a heart stroke in the future. The software will also generate comprehensive data visualizations that will help the field staff interpret the results easily.