The accuracy in this one seems very low, but the StrategyPage article emphasizes a mindset where some knowledge is better than none.

…an algorithm developed by an army reserve officer, 2nd lieutenant Christian Lance Relleve, whose academic studies covered architecture as well as HSGI (Human Security and Global Intelligence) and International Relations. Relleve presented an algorithm that could predict the internal layout of a building with 70 percent accuracy based on what country the building was in, what the apparent purpose was and obvious external features. Relleve noted that there were many external indicators of how the internal layout was and he examined layouts in many countries for various types of structures.

I’m kind of curious if the accuracy is even lower in reality because many of the “correct” assessments are just warehouses, toilets or similar single-use simple construction.