🧪 Biomedical Experimental Data Analysis Suite

An automated Python-based framework for statistical analysis and visualization of biomedical experimental data. This toolkit is designed for physiology, neuroscience, and general biomedical research datasets, enabling fast, reproducible, and publication-ready analysis from raw CSV files.

🚀 Overview

This project provides three levels of analysis pipelines:

1. Single-Parameter Analysis

• Summary statistics (mean, SD, n)
• Group comparisons (t-test or ANOVA depending on design)
• Basic visualizations (bar plots with error bars)
• Simple reporting

2. Multi-Parameter Analysis

• Batch processing of multiple dependent variables
• Automatic group-wise statistical testing per parameter
• Individual plots generated per variable
• Combined reporting in a structured Word document
• Correlation analysis across parameters

3. Advanced Analysis Pipeline

A full statistical framework with publication-grade rigor:

🧠 Assumption Testing Shapiro-Wilk test (normality per group) Levene’s test (homogeneity of variance)

📉 Outlier Detection IQR-based outlier detection

📊 Adaptive Statistical Testing One-way ANOVA (parametric conditions met) Kruskal-Wallis test (non-parametric fallback) Automatic decision switching based on assumptions

📌 Effect Size Estimation Eta-squared (η²) Omega-squared (ω²) Cohen’s d (pairwise comparisons)

📈 Confidence Intervals 95% confidence intervals for group means

🔬 Post-hoc Analysis Tukey HSD test for multiple comparisons

📉 Visualization Suite Bar plots with CI error bars Boxplots with individual data points (swarm overlay) Significance annotations (*, **, ***, ****) Correlation heatmaps

📄 Automated Reporting Word document (.docx) generation Embedded plots and statistical tables Structured parameter-by-parameter reporting

📂 Project Structure

biomedical-analysis/
│
├── single_parameter_analysis.py
├── multi_parameter_analysis.py
├── biomedical_experiment_full_advanced.py
│
└── outputs/
    ├── plots/
    └── reports/

🧰 Requirements Install dependencies:

pip install pandas numpy scipy matplotlib seaborn statsmodels python-docx

📊 Input Format

CSV file should follow this structure:

Animal ID	Group	Parameter1	Parameter2	...
1	Control	5.2	10.1
2	Treated	6.3	11.4

• First column: Sample ID
• Second column: Experimental group
• Remaining columns: Numeric biological parameters

---

⚙️ How to Run

Single Parameter

python single_parameter_analysis.py

Multi-Parameter

python multi_parameter_analysis.py

Advanced Full Pipeline

python biomedical_experiment_full_advanced.py

---

📈 Outputs

Each run generates: 📊 Publication-quality figures (PNG) 📄 Word report with full statistical summary 📉 Correlation heatmaps (advanced mode) 📁 Organized output folders per dataset run

🧠 Key Features Fully automated statistical decision-making Handles both parametric and non-parametric data Publication-ready visual outputs Reproducible analysis pipeline Scalable to large experimental datasets Built for biomedical and neuroscience research workflows

---

🔬 Intended Use Cases Physiology experiments (glucose, insulin, hormones) Neuroscience behavioral studies Pharmacology dose-response analysis Metabolic and oxidative stress studies Multi-group animal experiments

---

📌 Future Improvements Machine learning-based pattern detection Mixed-effects modeling for repeated measures Interactive dashboard (Streamlit/Plotly) Integration with electrophysiology and imaging datasets Automated hypothesis generation

---

👩‍🔬 Author Developed by Mary Oluwatobi Dada, Physiology graduate at the University of Ilorin, Nigeria. Built to support reproducible statistical analysis in ggeneral biomedical experimental workflows. LinkedIn:www.linkedin.com/in/mary-dada-1b4664139 · Email: dada.mary14@gmail.com

---

📜 License

For academic and research use. Modify freely with attribution.