/some-common-statistical-methods

python module, showcasing computation (as part of a learning process) of some common statistical methods including mininum sample size, confidence interval estimation methods for mean or proportion, hypothesis testing mehods and regression models witth metrics and test suites

Primary LanguagePythonApache License 2.0Apache-2.0

Statistical Analysis Toolkit

Welcome to statanalysis, a repository of statistical methods and tools tailored for data analysis enthusiasts. Inspired by my completion of a Coursera certificate in statistics, this repository encompasses a plethora of statistical concepts meticulously crafted into implementations. From prediction metrics to regression analysis, hypothesis testing to confidence intervals, and population parameter estimation to model estimation, statanalysis covers it all.

Built in Python, statanalysis provides meticulously crafted modules and utilities aimed at beginners in statistics, data science, and research. While following a certification on statistics on Coursera, I chose to solidify my knowledge through implementations instead of solely relying on existing modules. I believe there is no better way to understand a statistical formula than by implementing it in code, documenting it thoroughly, and validating the results through tests.

So, I've rewritten common statistical learning tools then create a repository that offers direct access to my implementations, ensuring simplicity without compromising accuracy. Futhermore, these implementations have undergone rigorous testing against established libraries like scipy.stats, statsmodels, and scikit-learn to uphold industry standards.

I have uploaded this open source project to pypi python module available on pypi, documented on readthedocs.

Whether you're a novice or an experienced data analyst, statanalysis aims to simplify and enhance your statistical analysis journey. Dive in and explore a wealth of statistical methods and techniques designed to streamline your analytical processes and empower your insights.

Features

  1. Utility Functions:

    • Module: utils_md
    • Description: The utils_md module provides a collection of helper functions for various statistical tasks, including data preprocessing, standard deviation estimation, and computation of probabilities and percentiles.

  2. Hypothesis Validation:

    • Module: hyp_vali_md
    • Description: The hyp_vali_md module includes functions for hypothesis validation, such as checking residuals, coefficients, and conducting hypothesis tests. Features encompass:
      • Constraint Checking: Functions for verifying constraints, such as checking if values fall within specific ranges.
      • Hypothesis Sample Size: Tools for ensuring minimum sample sizes for hypothesis testing scenarios.

  3. Confidence Interval Estimation:

    • Module: conf_inte_md
    • Description: The conf_inte_md module offers methods for estimating confidence intervals for population parameters, such as proportions and means. Features include:
      • One-sample Proportion: Functions for estimating confidence intervals for population proportions based on a single sample.
      • Two-sample Mean: Methods for computing confidence intervals for the difference between two population means, considering paired and unpaired data.

  4. Hypothesis Testing:

    • Module: hyp_testi_md
    • Description: This module encompasses a comprehensive suite of functions for hypothesis testing, covering a variety of scenarios:
      • Testing Population Proportions: Methods for assessing hypotheses related to population proportions using z-tests.
      • Comparing Means: Functions for conducting hypothesis tests to compare means between two or more populations, employing t-tests and ANOVA.

  5. Model Estimation:

    • Module: mdl_esti_md
    • Description: The mdl_esti_md module houses classes and functions dedicated to model estimation. Notable features include:
      • Linear Regression: Implementation of linear regression models, including ordinary least squares (OLS) and robust regression.
      • Logistic Regression: Classes for logistic regression analysis, enabling binary classification tasks with probability predictions.
      • Multiple Regression: Tools for conducting multiple regression analysis, facilitating the exploration of relationships between multiple independent variables and a dependent variable.

Repository Structure

The repository is organized into two main folders:

  1. statanalysis/ Folder:

    This folder contains the following modules:

    • utils_md: Module for utility functions, offering a collection of helper functions for statistical tasks.
    • hyp_vali_md: Module for hypothesis validation, containing functions for checking residuals, coefficients, and conducting hypothesis tests.
    • conf_inte_md: Module for confidence interval estimation, providing methods for estimating confidence intervals for proportions and means.
    • hyp_testi_md: Module for hypothesis testing, including functions for conducting hypothesis tests on proportions and means.
    • mdl_esti_md: Module for model estimation, including classes and functions for linear regression, logistic regression, and multiple regression.

  2. tests/ Folder:

    This folder features tests for all methods mentioned above.

Usage

To utilize the statistical analysis functionalities provided by this library, you have either clone the repo or install from pypi depending on your usage

Clone the Repository:

Clone the repository to your local machine using the following command:

git clone https://github.com/hermann-web/some-common-statistical-methods

Install the Library from PyPI:

Install the library from PyPI using pip:

pip install statanalysis

Choose the option that best suits your needs and get started with your statistical analysis.

Import Modules or Functions:

In your Python script, import the desired modules or functions using the following syntax:

from statanalysis import utils_md, hyp_vali_md, conf_inte_md, hyp_testi_md, mdl_esti_md

Perform Statistical Analysis:

Utilize the imported functions and classes to perform a wide range of statistical analysis tasks on your data. For example:

# Example: Compute a confidence interval for a population proportion
confidence_interval = conf_inte_md.IC_PROPORTION_ONE(sample_size=100, parameter=0.5, confidence=0.95)

Leverage advanced statistical techniques and methodologies provided by the modules to analyze your data effectively.

Additionally, if you prefer to browse documentation in a more structured format, you can refer to the online documentations which provides detailed information about the library's functionalities and usage.

Additional Information

  • The repository includes a comprehensive test suite in tests folder to validate the accuracy and consistency of the implemented methods against standard industry-standard libraries like scipy.stats, statsmodels, and scikit-learn.
  • The module is available on PyPI for easy installation and use in various statistical analysis projects.
  • For detailed explanations and references, refer to the respective sections in the code files or read the documentation on readthedocs
  • Further insights and explanations on statistical concepts can be found in the provided links.
  • For inquiries or assistance regarding the repository, please contact Hermann Agossou.