Three layers, each with a different audience and a different standard of proof.
JaxOps, departmental scorecards, originally Excel files on SharePoint. JaxAnalytics, the cube layer, where the Finance cube was the showpiece. And JaxScore, the public-facing dashboard — 1.0 launched in March 2014, with 2.0 in development through 2014 and 2015.
The program was named a 2015 CIO 100 honoree by CIO magazine, announced on 30 April 2015, and the Jacksonville Business Journal’s coverage of the division’s Digital Cities ranking named the performance management system and JaxScore directly.
Why a public dashboard is a different problem
A public dashboard is not an internal report with nicer styling. The numbers are disclosed; they are a public record once published; and any resident can dispute them.
That constrains the model underneath rather than the presentation on top. An internal report that is wrong gets corrected in a meeting. A published figure that is wrong is a public record that was wrong, and defending it means being able to walk backwards from the number on the page through the cube, through the collection step, to the operational system that produced it. Everything upstream has to survive that walk.
The collection layer is what decides whether it is evidence
JMIS, the JaxScore Metric Input System, replaced the first step of the scorecard process in 2015. Managers had been typing values into Excel scorecards on SharePoint and then into further spreadsheets for container values; JMIS took step one, with a second phase intended to take step two.
Served as one of the manager volunteers on quality assurance for the February 2015 pilot — and the launch is worth recording for the wrong reason. An authentication prompt, then a page not found, with nobody certain whether the application was running on the development or the production business intelligence host.
That is the whole argument for environment separation on an analytics platform, delivered by example rather than by principle. Elsewhere in an estate, confusion about which environment a thing is running in is a hygiene problem. Here it is epistemic: when the people launching a system cannot say which environment it is in, what it produces is not yet evidence. A number is only as good as the ability to say where it came from, and that includes which host served it.
The entry is here with the failure in it because a performance management record that showed only the parts that went well would be the one kind of record this program existed to argue against.
Outcome
Named a 2015 CIO 100 honoree by CIO magazine, announced 30 April 2015.