If you've ever conducted a statistical analysis, you might’ve spent hours thinking about which variables to include and the impact each would have on the outcome. But to ensure the model itself is accurate, shouldn’t someone measure the measurers?

In the early 1950s, a young Japanese scientist named Hirotugu Akaike asked this simple but crucial question. More than two decades of research later, he presented the answer as a simple equation known as the Akaike Information Criterion. With AIC, analysts select a model from a set of options by measuring how close the results are to the (hypothetical) truth.

For Dr. Akaike, experience was core to creativity. To get ‘a direct feel of random vibrations,’ for example, he bought a scooter and rode it around Mount Fuji. This first-hand experience helped him differentiate between the vibrations of riding on normal and heavily-trucked roads. 

Today’s Doodle portrays Dr. Akaike against a Google-inspired approximation of functions, parameters, and their respective curves. 

Below are a few initial conceptualizations of the Doodle.