Comic Book Lessons in Analytics — Loki
Defining a Single Source of Truth in a Multiverse
This article is about the Marvel Studios series Loki, currently airing Season 2.
The first season of Marvel Studios Loki introduced us to the Time Variance Authority (TVA), a bureaucracy running in the shadows of reality to protect what it called “the sacred timeline”.
The sacred timeline is, in simplest terms, the “correct” version of all events happening as they are supposed to happen in order to protect time itself. The TVA accomplished its goal by pruning anyone or anything that threatened the sacred timeline, removing them completely from existence.
It is a very “comic book” plot. It is also a perfect analogy for a discussion about a single source of truth (SSOT) in data.
Sacred Timeline
The sacred timeline was defined in the first season of Loki as the correct version of events, to be protected at all costs. Without the constant vigilances of the Time Agents, branching timelines would run amok.
Loki, as the titular hero, was told that any branching was dangerous but movement outside the red lines on the monitor would be a disaster. He was also engaged to help capture a multiversal variant of Loki (Sylvie) that was causing all sorts of trouble.
Loki, like most us, did wonder just who defined which version of events was the version to be protected. Sylvie, the variant being pursued by the TVA, was wondering the same thing.
Spoiler alert: the “sacred timeline” was the version that one person liked.
This individual introduced himself as “He Who Remains”. Astute viewers, comic readers, and anyone who saw Quantumania would also recognize him as Kang (the Conqueror), also known by many other names throughout history.
Kang is complicated. Rather than attempting to explain, I will summarize. In the comics, he was first encountered in ancient Egypt as the Pharaoh Rama-Tut, though he originated in the 30th century, and it just gets more complicated from there.
Kang has had many names. His original name is / was / will be Nathaniel Richards, named for his 20th century ancestor who is the father of Reed Richards of the Fantastic Four.
Pruning the Timeline
Obviously, you will not be in the position of pruning aspects of reality in your analytics work. I hope. But the analogy, while heavy-handed and a bit on the nose, is appropriate.
Kang (or He Who Remains, or whatever you prefer) had to ignore and destroy more data than he kept in order to dictate his single source of truth, his “sacred timeline”. It was not accurate, and it was fragile.
Throw a few more Loki variants into the mix, and his control started to fail.
OK, I admit that I just wanted to use that picture and that caption.
In analytic terms, Kang was attempting to use a specific set of data in order to create the results he wanted in the first place.
With the technology at his disposal, he likely could have put a system in place to catalog the real collection of events (the multiverse).
We have already seen the temporal loom in the second season opener.
The real shame is that Loki, the Norse god, saw this loom and made no reference to the Norns. I mean, it was right there.
Single Source of Truth
No data or analytics architecture will be useful without a single source of truth (SSOT). While I wouldn’t use the word “sacred” to describe it, the SSOT must be accurate, correct, and, as much as possible, complete.
Without that source of truth, we cannot claim to have accurate data.
We do need to decide which system or database is the SSOT, or we need to create one. We do not, however, manipulate the data capture or curation in order to affect the results of our analysis / data science / AI models.
Where the SSOT is incomplete, we need to document it clearly and work to remediate any gaps. We cannot just ignore the data that we do not like or that is difficult to reconcile (the joker data).
That data exists for a reason, and it is important to understand it.
Keep it accurate. Keep it honest. And keep improving.
For more perspectives on truth in data & analytics, check out:
