This project demonstrates the application of exploratory statistics techniques to analyze differences between groups in various experimental scenarios. The project uses Analysis of Variance (ANOVA) and Multivariate Analysis of Variance (MANOVA) to explore the data and visualize the results.
The data used in this project is synthetic, generated to simulate different experimental conditions. Each dataset represents a unique experimental scenario with varying numbers of groups and dependent variables.
- First Dataset: Simulates an experiment with five groups at different levels of a given condition.
- Second Dataset: Represents a hypothetical experiment involving three groups with two dependent variables.
- Third Dataset: Mimics a 3x3 experimental design, with nine total conditions and with three dependent variables.
The project employs a variety of statistical techniques and data visualization tools to analyze and interpret the data:
- ANOVA: Used to analyze the first dataset and identify significant differences between the groups.
- MANOVA: Applied to the second and third datasets to analyze the data and identify group differences.
- PCA Biplot: Used to visualize the group differences in the second dataset.
- 3D Plot: Utilized to visualize the results of the third dataset, representing the three independent variables.
The analysis revealed meaningful differences in at least one condition for each hypothetical scenario. In each case, the null hypothesis would fail to be rejected, indicating significant differences between the groups.
The project is written in Python and uses several libraries, including pandas, statsmodels, matplotlib, seaborn, numpy, sklearn, and plotly. To reproduce the analysis, these libraries should be installed in your Python environment.
This project showcases the application of ANOVA and MANOVA techniques in analyzing group differences, demonstrating the ability to interpret complex datasets and draw meaningful conclusions from them. It also highlights the use of various data visualization tools to effectively represent the results of the analysis.