Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/gyakobo/covid-19-case-study
The project performs data analysis and machine learning modeling using a Covid-19 dataset from New York City. The purpose is to analyze trends in Covid-19 cases, deaths, and hospitalizations, and predict hospitalization counts using regression models.
https://github.com/gyakobo/covid-19-case-study
covid-19 k-means linear-regression njit pandas python
Last synced: 10 days ago
JSON representation
The project performs data analysis and machine learning modeling using a Covid-19 dataset from New York City. The purpose is to analyze trends in Covid-19 cases, deaths, and hospitalizations, and predict hospitalization counts using regression models.
- Host: GitHub
- URL: https://github.com/gyakobo/covid-19-case-study
- Owner: Gyakobo
- License: mit
- Created: 2024-12-16T19:57:26.000Z (about 1 month ago)
- Default Branch: main
- Last Pushed: 2025-01-05T07:04:08.000Z (11 days ago)
- Last Synced: 2025-01-05T08:17:17.713Z (11 days ago)
- Topics: covid-19, k-means, linear-regression, njit, pandas, python
- Language: HTML
- Homepage:
- Size: 366 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Covid 19 case study
Author: [Andrew Gyakobo](https://github.com/Gyakobo)
https://github.com/user-attachments/assets/617cf83f-d8c9-4cba-b72a-4892b3620803
## Introduction
The project performs data analysis and machine learning modeling using a Covid-19 dataset from New York City. The purpose is to analyze trends in Covid-19 cases, deaths, and hospitalizations, and predict hospitalization counts using regression models.
## Project Breakdown
1. Task 1: Exploratory Data Analysis (EDA)
* Data Source: JSON dataset retrieved via an API endpoint.
* Data Cleaning:
* Dropped rows with NaN or empty values.
* Converted date columns to datetime and extracted the month.
* Visualization:
* Line plots of cases, deaths, and hospitalizations over time.
* Histogram of daily Covid-19 case counts.
* Boxplots showing case distribution across months.
2. Task 2: Linear Regression Model
* Goal: Predict hospitalization counts.
* Steps:
* Preprocessing:
* Features selected include month, case count, and death count.
* One-hot encoding for the month feature.
* Split data into training and testing sets.
* Trained a Linear Regression model.
* Evaluation:
* Measured model performance using:
* Mean Squared Error (MSE)
* R² score
* Visualized predictions against actual hospitalization counts using scatter and line plots.
3. Task 3: Logistic Regression for Classification
* Goal: Classify hospitalizations as "high" or "low" based on a defined threshold.
* Steps:
* Thresholding on hospitalization counts (e.g., 60th percentile).
* Trained a Logistic Regression model.
* Evaluation:
* Accuracy score and classification report (precision, recall, f1-score).
* Confusion matrix visualized using a heatmap.
## Key Results:
* Linear Regression:
* Provided a good fit with R² ~ 0.77 and a clear relationship between the features and hospitalizations.
* Logistic Regression:
* Achieved an accuracy of 85.5%, effectively classifying "high" and "low" hospitalizations.
* Confusion matrix shows a balanced prediction performance.
## Subtle Description
This project efficiently combines exploratory data analysis and machine learning to analyze and predict Covid-19 hospitalization trends. By leveraging statistical models like linear regression for prediction and logistic regression for classification, it provides insights into the relationships between case counts, deaths, and hospitalizations. The clean visualizations, such as time-series plots, histograms, and confusion matrices, make the findings intuitive and actionable.
## License:
MIT