Kanban Metrics

This is an experimental Jupyter notebook and command line tools using Jira issue changelogs to:

1. load a historical record of a Kanban
2. summarize key Kanban metrics
3. graph those key Kanban metrics (cycletime, throughput, flow, etc)
4. use these metrics to run a Kaplan-Meier survivability model and a Monte Carlo simulations

This allows us to use data forecasting (instead of gut estimation) to gauge:

• how likely will a single issue be completed in N days
• how much work a team can take on over the next N days
• and by what date N number of work items could be completed

Setup

pip install -r requirements.txt

Fetching Data from Jira

python jira.py [project_key] [since_date]

Analyzing Data from the Command Line

python analysis.py -f data/data.csv summary

Analyzing Data from Jupyter

jupyter lab

Examples:

Generated from an example Jira project and data loaded from example.csv on 2020-08-25 covering the previous 90 days: