Pychometrics Python Library

Item Analysis Simplified

Pychometrics is a simplified item analysis python library that analyses assessment and questions (items). It gives straight forward interpretation based on the calculated parameters.

Features

• Read csv file containing students scores
• Caclulates measures of central tendency of the assessment
• Aslo computes coefficient of internal consistency (Cronbach Alpha) to assess the consistency and reliability
• Item analysis for each question through Facility Index and Discrimination Index
• Item level feedback based on these parameters

Dependencies

List of dependencies required to use your package: - pandas~=2.2.2 - setuptools~=68.2.0 - reportlab~=4.2.0 - numpy~=1.26.4 - scipy~=1.13.1

Online Demo

A minimal working online demo is available here

Installation

Pychometrics is a python library and requires latest python library (tested in python 3.11).

Install the library with the dependencies through github or PyPI.

```sh pip install -U git+https://github.com/Shiva-DS24/pychometrics.git@main

or from pyPI: pip install pychometrics==0.1.4

from jupyter: !pip install pychometrics==0.1.4

``` After installation, prepare a score data file as per the template provided here. Please note that this is similar to Moodle quiz export responses file and therefore that file can be used as it is. If assessments are conducted outside any Learning Management System, please prepare a csv file having column header Question No or Name with '/Max Marks'. Please refer the sample data file provided here.

Import the library sh from pychometrics import AssessmentAnalysis, generate_pdf_report Then read the data file and run the analysis ```sh

read data file

analysis = AssessmentAnalysis('data.csv')

run the analysis

analysis.run_analysis() ```

More functions:

1. If the data file is in the working directory and named as ‘data.csv’, data can be read using: analysis = AssessmentAnalysis('data.csv'). once read the message will indicate the number of students and items (questions) identified along with the missing values if any.
2. To get information about the loaded data at any time, we can run analysis.info().
3. The descriptive statistics of the entire test can be obtained by running analysis.calc_desc().
4. The skewness and kurtosis can be calculating by running analysis.calc_skew() and analysis.calc_kurt() respectively.
5. The cronbach’s alpha can be obtained by running analysis. calc_alpha().
6. The Standard Error of Measurement (SEM) can be calculated using analysis.sem().
7. Calculation of FIs and DIs for each item can be obtained using analysis.calc_fi() and analysis.calc_di() respectively. Further to this analysis.indices_export() export the data as a csv file in the working directory.
8. A comprehensive pdf report can be generated using analysis.export_report() command.

9. To obtain help about this library: analysis.help(), this will display available functions and their syntax.

   The generated pdf file will be available in the working directory.

Contributing

If you would like to contribute to this project, you can fork the repository and send us a pull request. We welcome contributions!

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Support

If you have any questions, issues, or suggestions regarding this package, feel free to create issue here.