Superhero Lessons in Analytics — Ep. XIII
Defining Resolution — Lessons from Ant-Man
Resolution is not a common analytic phrase. It isn’t much talked of outside photography and chemistry. But then, Paul Rudd is not a common choice for a superhero either. If Marvel can try new things, so can we. Maybe it is too late to change the name… but I am going to try anyway.
For the record, Ant-man was silly. I never read the comic book. I don’t know how accurate it was. What I know is that I kept waiting for Rick Moranis to show up and wondered if Anty and Anthony were played by the same actor. I also laughed and thoroughly enjoyed it. Superhero movies are allowed to be dumb. Analysts… should avoid that sort of thing.
Resolution
How big? How small? Ant-Man dealt a great deal with resolution. Only the movie kept getting confused. The scientific mumbo-jumbo the audience was provided went something like this: the PIM particle allows the space between atoms to be contracted (shrinking) or expanded (giant toy trains). Thus little Ant-Man has the strength of a full-size Ant-Man. So how was Henry carrying around a shrunken tanks???
Analyst get confused, too. Data mining is the more common term thrown around instead of resolution. Alternately, many ask “what level is this data?”. A typical answer is either account level, customer level, transaction level, or segment level.
The government even creates rules in the US preventing the use of analysis at certain levels. In banking, there are constraints for using zip + 4. In photogrammetry (aerial map making), our satellite imagery was often purposely distorted because our government clearance was not high enough. Seriously, blurred maps for the military… brilliant! Of course the government is always confused…
Ant-Man’s writers fared no better. They actually based their plot around the danger of getting too small but then failed to account for things like Thomas the Train. The entire toy battle was silly. None of those toys, normal-sized or enlarged had the weight and density to hurt anyone. Thomas should have exploded into dust particles or something more logical and pseudo-sciency.
But then, Ant-Man fails at every level if real science is brought to bare. If one goes subatomic, how can you still see?
It does succeed at providing a warning. Unless you are aiming at writing a comedy, you need to be wary of the resolution of your data. From simple axis resolutions in graphs that insinuate distorted trends, to failures to account for population sizes when drilling down into the data, all these run the danger of destroying your work and wasting your time. They wreak havoc on averages and time series as well.
Also note, just like Ant-Man, this stuff can bleed into other analysis (or movies). In Avengers Civil War, giant Ant-Man would have still weighed the same as normal Ant-Man. Of course Scott was comic relief for a story that needed some… I doubt your next analysis is looking for some carry over comedy.
Thanks for reading and be sure to check out the rest of the series.