During an air flight, the pilot can switch to 'autopilot' mode to rest his attention; if he is equipped with an ILS (Instrumental Landing System), his plane can even fly, land and taxi itself to the parking lot without any intervention from the pilot... provided that the airport where he lands is indeed equipped with adequate radars and that his ILS system is not out of order... In short, in 99,...9% of the cases, things happen as expected, and human intervention is superfluous. The remaining 0,...1% of the cases where it is essential.
The same is often true of computer systems. The processes work and meet the specified needs until the unexpected happens and requires an exceptional manual handling, in "fireman mode". Naturally, we try to improve the system and we usually succeed...but this does not prevent another malfunction or unexpected behavior from occurring one day. In short, any complex and evolving system is doomed to the myth of Sisyphus... That's why there are still pilots in airplanes and drivers in autonomous cars.
In Business Intelligence, dashboards are also complex and evolving systems. The challenges, constraints and expectations of business departments in terms of key process indicators (KPIs) are evolving more and more rapidly, requiring ever greater responsiveness and agility from IT teams in updating these indicators.
The dashboards developed with perseverance by the IT teams respond at best to 99.9% of the standard requests; but there will always be a need for a new indicator missing from the usual reports on the day, to explain a situation or justify a decision. And there, we know how it ends: in "system D" mode with a good old export in .CSV from the most updated production database, and a processing in Excel bodybuilded by magic macro.... In short, the whole decision-making edifice and the quality of the processing built with application over time are called into question in a hurry... and all this for a good cause, of course.
There is no miracle solution, but there is a path worth exploring to try to solve this headache. If the data sources and their preparation processes require stability and total control from the IT teams (to guarantee their reliability, quality and consistency with the information system), the same cannot be said for the dashboards themselves: they should remain the prerogative of the business departments, who alone know which key indicator, which KPI, is useful on the day and how to make it readable through a relevant dashboard that they have updated themselves, without the intervention of the IT teams, and without delay or complicated processes...
This is what Data Visualization is allabout: guaranteeing the content (the data) while controlling the form (the information) through dynamic and 'intelligent' reports designed and adapted by the Business Departments in an autonomous manner. To go from Excel to Business Intelligence...
