Consider Seurat's Grande Jatte. Since artists were beginning to consider themselves intellectuals, the paining was intended to show the banality of bourgeois life (in this case, the Sunday promenade).
|Sunday Afternoon on the Island of La Grande Jatte |
(Un dimanche après-midi à l’Ile de la Grande Jatte),
Georges Seurat, 1884-1886.
But what it does show is technique. From a certain distance, we see a solid picture. But now focus closely on the dog. Look at the location just above the dog, at the edge of the shadow. Look closely.
|Grande Jette, detail edge of shadow above the dog.|
and what was one a solid picture resolves into teeny points of individual color.
So what has this got to do with statistics?
Here is a long-view of temperatures and carbon dioxide running back to nearly 450,000 years before present (ybp).
Red is CO2 and blue is temperature in °C. The horizontal line is the baseline temp for 1960-1990 used in
Here is a closer look at a dot: Greenland ice core data for only 11,000 ybp:
The upper chart is temperature and shows (vertical green stripes) the Minoan, Roman, Medieval, and Modern Warm periods. In between are the collapse of the pristine states, the Dark Ages, and the Little Ice Age. Interestingly, except for the Warm Period spikes, the overall trend since Minoan times has been down. Each Warm spiked cooler than its predecessor; except possibly the current one.
The bottom chart is atmospheric CO2 from the ice cores. This began increasing in what looks to the eyeball like a logistic curve starting about 7000 ybp -- at the same time the temperatures were cooling off. There is no appreciable correlation with the Warm spikes.
A closer look still -- at 32 ybp scale -- and the temps once more seem to match the CO2, if you discount the current flattening. These are the usual charts you see presented.
The question arises: which scale is the right scale for considering these things? Look too close and random fluctuations loom as stupendously important; look too broadly and you see canals on Mars.