TRACC - Tracking Record of Aggregate Changes over Commits

Julian Eder, Justin Dachille

Motivation

Git Version Control

• Allows for complex development
• Git servers and platforms
• Insights into codebase and team dynamics are valuable for project managers

https://blog.cpanel.com/git-version-control-series-what-is-git/

https://github.com/opencv/opencv

Visualizations on Github

• “Github Insights” - limited

Related Work

JMRI & Gource

• Some related visualization tools

https://www.jmri.org/community/visuals/codeswarm.shtml

https://gource.io/

Objectives

Target Insights

• Which parts of the code did development focus on, at different points in time?
• Which contributors added the most/least code to the project? (as a basis for personnel decisions, to reveal the code ownership across different parts of the project)
• Identifying which contributors are experts for a particular part of the code? (as a basis for distributing future development tasks)
• Which files or modules have the highest rate of frequent changes?
• How has the codebase evolved over time?

Project Overview

• Utilize Flask, Python, D3 to display data from Github REST API

​

Visual Analytics Design

• Overview of project file tree
• Contributions of Users per file
• Contributions of selected commit visualized as links to files
• Analysis Methods
• Scroll through repo history
• Inspect individual files and contributors (links to github for further drill-down)
• Scalability:
• This is meant for managers of small to medium sized teams
• Visual scalability for pie diagrams could be a problem

​

Mockups

• Activity by file/folder over time

​

• Develop a sense for when a challenge has gone on too long

​

• Example cases:
• Find where development time might be wasted
• Find where certain devs are struggling

​

Provisional: Code Churn

• Extension of previous visualizations, if time permits

Challenges

• Visual Scalability
• Visualizations with dynamic data
• Ex: Runtime

Questions?