Write Effective Code

What is “effective” code? The notion of effectiveness evolves as a data science project moves through the develop, deploy, and scale stages.

In Development, effectiveness is measured by the data scientist’s ability to use code to solve a problem. Code that is burdensome to maintain or not optimized to perform under demand is not effective.

The articles below offer guidance through these stages.

Featured

Minimum Viable Python — Working with Python

These articles contain guidance on how to incorporate Python into your data science workflow. They are not intended as general tips for all Python users, but rather as an…

Learning to code in R and/or Python

When just beginning, these resources can help

Training

Posit Academy
The most effective way to learn data skills with your team

Getting comfortable with your tools

Minimum Viable Python

Using Python with Posit Products — an Overview

Connect User Guide

Workbench User Guide

Examples

Moving from Excel to R
See insights from an Excel workbook extended with code-based outputs such as Shiny, Flexdashboard, and R Markdown

Using `reticulate` to combine R and Python in one project

Learn new frameworks

Whether R or Python, there are new tools to add to your toolbelt

Analyses and Reports

Add interactivity or parameterization to your reports for greater flexibility

Quarto for R, Python, Julia, and Observable

R Markdown

Jupyter

Web Applications and Dashboards

Without knowing html, javascript, or CSS, you can create powerful interactive applications and dashboards

Shiny for R and Python

Dash

Streamlit

APIs

Use APIs as a means to expose your analytics to other systems or create pipelines

Plumber

Flask

FastAPI

Example: integrate a plumber API with Slack

Example: expose a ML model as a plumber API

ModelOps

Vetiver for R and Python

Example: Bikeshare project model deployment using vetiver

Code Smart

Be efficient with your code and your time.

Parameterization

Create a report template and customize your outputs with parameters

Learn about Parameterization with R Markdown

See example gallery projects using parameterized R Markdown

Leverage Different Output Formats

Example: Create and Publish a Slideshow from Jupyter Notebooks

Example: Access to Care

Break out work into reusable pieces

Example: Repeatable Data Science with the Pins Package

Utilize Automation

Send customized or conditional emails from scheduled reports on Connect

Utilize `targets` for orchestration

Deploy Responsibly

Shiny Feature Toggles
Feature toggles can be used to deploy changes to production (new features) that are hidden to most/all users and then be tested under production load and incrementally exposed to user groups as your confidence increases.