The real nub of the issue is this: in the absence of data, we can only use data that is present. Here, it’s clear when someone commits a crime and we catch them after the event with a certain number of mobile train security personnel on our payroll. That’s measurable: the ratio of events to arrests, for example. What’s not measurable by automated data science and analytics half so easily is when something doesn’t happen because a permanent guard is present to act as deterrent.
Mil Williams, Stockholm Sweden, 21st April 2023
Introduction:
There are strikes on the commuter trains — the otherwise fabulous pendeltågs — here in Sweden: even the occasional wildcat ones. The frustration is patent: more so, because the strikers are right.
This is why:
An aside:
As a brief by the by before I continue, I think the train companies are able to claim the numbers of security and safety staff would remain the same, and yet still want to go ahead with it all, because they’re changing the type of workforce: you still need to go through with rightful and rigorous measures to vet and upskill non-train guards of all sorts it’s true, but with a train guard it’s less easy to change and chop their working locations, conditions and so forth. Or outsource the workforce, even. Change overnight who employs them and how.
No?
So …
How a data-driven world can deceive:
The thing is, here we have a perfect example of when a “data-driven world” actually needs academia more than it needs an automated data analytics and data science as we usually understand them.
The train companies in Liverpool and Sweden both I am sure will have had long-term strategies to re-engineer the structures of their employees and related re in-house and outsourcing options, and whilst taking guards off the trains in the circumstances described wouldn’t deliver immediate economic advantage, as indeed they underlined in Liverpool for sure, long-term if I’m right it definitely would.
The real nub of the issue is this: in the absence of data, we can only use data that is present. Here, it’s clear when someone commits a crime and we catch them after the event with a certain number of mobile train security personnel on our payroll. That’s measurable: the ratio of events to arrests, for example. What’s not measurable by automated data science and analytics half so easily is when something doesn’t happen because a permanent guard is present to act as deterrent.
And this is the challenge here. It really is a challenge around what we do when the evidence base is incomplete: that is, how it leads us to take quite the wrong decisions.
To the solution:
There is a solution too; I alluded to it above. Straightforward academia gives us tools to codify absences, in for example qualitative data such as an interview transcript or video, so that what isn’t said is as significant as what is.
If we could create an equal set of tools for strategic decision-making when deciding if to take train guards off trains or no, perhaps we would avoid the strikes we’re having everywhere: and at the very least, we could validate, in a less conflictive way, the common sense most users of public services have that a “bobby on the beat” engenders an incomparable feeling of safety even where a car in the neighbourhood can be evidenced to deliver on objective data relating to quantitative crime events.
Summarising:
In crime and public safety, what doesn’t happen is as important as what does: and the “why” of both these matters, too.
So.
Let’s do something after the evidence bases for both aspects of the truth: that which has a visible side and the invisible events as well.
And then let’s achieve delivery of these aspirations sooner rather than later.
Further reading:
https://en.wikipedia.org/wiki/Sweden_Democrats
Another by the by: the promoters of today’s information are a further example of why we should act on the basis of what is not visible, as well as what is.
The Sweden Democrats started out as fascist and redolent of nazism of the very worst sort — at least according to the English version of Wikipedia. They themselves claim to have re-engineered their political DNA, which is not impossible but highly unlikely. Even so, medical professionals claim bespoke DNA of the human kind is very close to becoming a reality now; so we could argue that in politics it’s not unthinkable any more.
Let’s just say, however, for the moment unlikely and hard to do.
So. The risk from relying on present datasets instead of datasets relating to both what’s present and absent too? We allow people to hijack in bad faith what needs to be promoted in good faith.
The train personnel are right. Guards on trains deliver safety and security. This Swedish political party — in the current security conditions which China and Russia together have been stealthily laying out for decades together — are also correct to highlight the dangers of such, separate, narratives.
But they are wrong to a) conflate two issues like this; and b) lever the abuse and violence of both nation-states and their outliers in the fields of geopolitics to then promote an immigration narrative of their own re Sweden which delivers total obfuscation of our all too human reality and a zero confusion around their racist truths. Unless you choose to remain confused.
Sometimes it’s right to be firm: China — not all Chinese people — is a toxic regime. Putin’s Russia, too, has absolutely no redeeming qualities. But firm doesn’t mean we have to give fascism a place at the table of a wider collective progress.
Don’t besmirch the truth of the train staff by taking political shortcuts. And if this is what changing your political DNA leads to, change is what clearly you are NOT delivering.
Just occurs to me, too: even more reason to proceed with #intuitionvalidation.
mil's page 