The next story is a morphed version of a real life one.
Here we go:
There is this process, the evolution of which can be tracked with one parameter, call it P.
According to science, P has a normal low value, say Lo and a normal hi value, say Hi.
The process is carefully being tracked, and after a short period of time P seems to break through the Hi value barrier and looks like set for a journey to the stratosphere.
Here’s the diagram.
“How far must this thing go before I should consider running out of here?”, is the question, gently put to the process specialists by the engineer.
And quickly came the answers:
“OMG, this is disaster, we need to stop this.
No, yes, maybe.
I don’t know. OMG. OMG!”
This was the panic stage.
Now we go for the zen status
Take a deep breath, do some research, and find out that there are other limits to P, limits that “specialists” obviously have forgotten:
A1 = 2.5 Hi, mild problem
A2 = 5 Hi, moderate problem
A3 = 20 Hi, severe problem.
Anything above A3: start running.
So, with the time-series at hand, and the knowledge about the levels, and a few lines of code, Wolfram Language by the way, the engineer produced this diagram.
And zen broke out.
Indeed, the forecast – black dotted line – and the 96% confidence band at the end of the planned process run, indicated that P would stay well within the “moderate problem” band, and nothing was to be feared.
How many times were you and team in panic state, where zen should have been the case?
More dangerous: How many times have you and team experienced zen, where you should have been, well, at least concerned?
- A good understanding of the process,
- A good dose of data science,
- And a few lines of code – given the right tools :
The best recipe yet available for permanent zen.