Rush Regression Workbench

Python source code for a plotly / dash app used for conducting various forms of regression.

Description and functionality

The Rush Regression Workbench is a freely available and lightweight dashboard application for conducting various regression analyses. This open-source tool automates iterative linear and polynomial regressions, multivariable forms of linear and logistic regression, linear and polynomial forms of quantile regression, as well as tests of regression assumptions, automated data encoding, optimized data transformations, machine learning based feature elimination, outlier detection, and construction of residuals plots, confidence intervals, prediction intervals, and diagnostic curves. The regression workbench attempts to recognize invalid data features, helps prevent invalid analyses, and offers tool tips and analytical interpretations.

Upload Data

Conduct iterative pairwise regressions

Conduct a single regression for deeper insights

Conduct linear and polynomial quantile regression

Conduct multivariable linear regression

Conduct multivariable logistic regression

More to come!

How to run this app

1. After downloading or cloning this repository, open a terminal window or command prompt in the root folder.

2. Create a virtual environment for running this app with Python (>=3.8)

   In MacOS or Linux:

   python3 -m virtualenv venv
   
   

   In Unix:

   source venv/bin/activate
   
   

   In Windows:

   venv\Scripts\activate
   

3. Install required packages using pip:

   pip install -r requirements.txt
   

4. Run this app locally with:

   python3.8 app.py
   

   The output of the terminal window will look like:

   Dash is running on http://0.0.0.0:8050/
   

5. Paste the url into your web browser and voila!

Requirements

These are automatically installed when following the instructions above.

• python version 3.8 or newer
• dash==2.0.0
• gunicorn==20.1.0
• numpy==1.22.1
• pandas==1.4.0
• scipy==1.7.3
• flask==2.0.2
• plotly==5.5.0
• datetime==4.3
• statsmodels==0.13.1
• scikit-learn==1.0.2
• dash_bootstrap_components==1.0.2
• lxml==4.8.0
• werkzeug==2.0.2
• jsonschema==3.2.0

Files & Directories

app.py

The primary file for running the Rush Regression Workbench application. This file contains the entirety of source code for the app as well as many comments to explain the application's functionality.

assets

Files in this directory are used by the application to format its interface or are used as images in this README file. All files except `RUSH_full_color.jpg` and those in the `images_for_README` were obtained from another open source Plotly Dash app (https://github.com/plotly/dash-sample-apps/tree/main/apps/dash-clinical-analytics/assets.): `Acumin-BdPro.otf`, `base.css`, `clinical-analytics.css`, - `plotly_logo.png`- `resizing.js`

• Acumin-BdPro.otf: An OpenType font file used by the application.
• base.css A cascading style sheets (CSS) used by the application. CSS is a stylesheet language used to describe the presentation of a document written in HTML or XML.
• clinical-analytics.css An additional css file.
• plotly_logo.png
• RUSH_full_color.jpg
• images_for_README: A directory containing png files used in this README document.

Procfile

This extensionless file is necessary for deployment on Heroku, and essentially tells Heroku how to handle web processes using the gunicorn server. The file contains a single line with the following: `web: gunicorn app:server`

requirements.txt

This file lists all of the software libraries needed for the app to run. When deploying the app on Heroku, this file is used to set up the server with the libraries necessary for running the application. When used locally, this file tells pip which libraries need to be installed (i.e., `pip install -r requirements.txt`).

runtime.txt

This file is used when setting up the app to run on an online Heroku server. It contains a single line: `python-3.8.16, indicating the version of python to use.

Contributers

• Developer: Kenneth J. Locey, PhD. Senior Clinical Data Scientist. Center for Quality, Safety & Value Analytics. Rush University Medical Center.

• Tester: Brittnie Dotson, MS. Clinical Data Scientist. Center for Quality, Safety & Value Analytics. Rush University Medical Center.

• Tester: Ryan Schipfer, MA. Senior Clinical Data Scientist. Center for Quality, Safety & Value Analytics. Rush University Medical Center.