Visual channels, and how accurately each one reads
The same eight market shares, drawn four ways. Pick any two territories and compare the true value ratio with the ratio the channel actually draws. Only one of the four keeps them equal and is also read accurately.
True value ratio vs drawn ratio1.15× vs 1.15×
Larger Riverina 23.0%Smaller Kimberley 20.0%Reads like about 1.10×
Shares are illustrative and add to 100.0%. "Reads like" applies the power law this concept states, with an exponent near seven tenths for area and near one for length. It is an average reading across many people, not a prediction about one reader.
Encoding
Area is drawn correctly here. Radius scales with the square root of the value, so the radius ratio is √1.15 = 1.07 and the drawn area ratio comes back to 1.15×, equal to the value ratio. The fault is in the reading, not the drawing. Area is under-read, so this gap looks like about 1.10×. A pair fifteen per cent apart, such as Riverina against Kimberley, ends up looking nearly identical.
Assigning channels by job. For monthly revenue across four product lines, revenue goes on vertical position against a common scale, month goes on horizontal position, and the product line goes on colour hue. Hue is an identity channel, so it names the series and never carries the quantity. Four separate lines do not draw the total, so a total needs its own panel on the same time axis.