OrionAI now measures your data before it draws the map

Two maps of the same numbers can tell different stories. Group population density in equal steps and almost every county lands in the palest colour, because one capital city stretches the scale. Group it by the natural gaps in the data and the differences between all the other counties appear.

OrionAI now measures before it decides. For every classed map it tries several ways of grouping the values, at three to seven classes, and scores each on how much of the variation it captures and whether every class is used. It keeps the best and tells you why in one line on the result card. Very uneven data, such as city sizes, gets methods built for long tails. A handful of areas gets fewer classes.

A colour key you can trust. The key now shows the real lowest and highest value in each class. A class that holds one area says so, for example "6,750 (1 area)", instead of an open-ended "above" row.

Counts or rates? Before styling a field it is unsure about, OrionAI looks at its values: is it a total that needs dividing by area or population, or does it run either side of zero and need a two-sided colour scale centred on no change?

A check before saving. Every saved map is checked for common cartography mistakes: colours that would not match the key, raw counts coloured on areas, rainbow scales on ordered data, empty classes and too many classes. OrionAI fixes what it can before it finishes.

You can still set the method, the number of classes or the colours yourself, in plain words.

Read the full OrionAI 2.14.0 release notes.