3 Tips for Better Data Analysis Using Strategies from Software Development

Similar to how developers produce software, analysts produce data products

Data products include dashboards, analyses, and data apps

1. Borrow from Agile

Agile is a process framework that focuses on iterative development done in short bursts, called sprints

Planning and prioritization at the start of each sprint

Use regular scrum meetings to conduct sprint planning, prioritization, and review

Clearly defining tasks with deliverables and timelines

Present hypotheses, approach(es), and aspects of data product(s)

Retrospectives and demos at the end of each sprint

Present analyses/data products to stakeholders, review sprint

2. Use READMEs

A README is documentation that tells you how to start using and understanding a new piece of software

A typical README is a text or markdown-formatted file included with the project

Learn how to format a README here: https://www.makeareadme.com/

READMEs for data products include things like usage, reproducibility, sources, and metadata

3. Integrate User Experience Testing

For a widely-used data product, consider User Stories

User Stories are a prioritized list of tasks the user is trying to accomplish

User Stories help developers, analysts understand users’ needs

Iterative, human centered design

Is the analysis interpretable? Is the way it’s presented intuitive and simple?

Prototype analysis with user groups, have them ask questions, document findings

Tips for Better Data Analysis

  1. Borrow from Agile
  2. Use READMEs
  3. Integrate User Experience Testing