Peter Drucker reportedly said “What gets measured gets improved.”
At work I was challenged to better track planned versus actual at the release and sprint level. My colleagues have gone down the route of using PowerBI to provide a summary of all the items in scope, the items with incorrect information, and a rolled-up view of the status and capacity used.
After some (brief) deliberation I decided to use Python instead of going down the PowerBI (or Tableau). My reasons were based on the fact that Python:
- Allows for dynamic data updates via Python JIRA module (or Rest API)
- Allows for customised output using TKinter
- Allows for the report to update JIRA
The setup that we currently have access to with PowerBI is updates can only be requested every 15 minutes, has a limited set of widgets, and only allows reading from JIRA (and not writing).
After tinkering for a couple of hours to get the Python code to return some data using the JIRA python library, I was then able to use LLM (Co-Pilot) to quickly iterate over the code to add support for extracting the required items, linking from the child issues to the parent issues (we use JIRA in a non-standard way), and even display some initial data using TKinter.
The main advantage of using the LLM is that I can iterate quickly. The main disadvantage is that it is generating code that I do not wholly understand, in particular the TKinter code.
