Testing Benford’s Law

There is an odd numeric phenomenon in large datasets. If a set of values were truly random, each leading digit (1 thru 9) would appear about 11% of the time, but Benford’s Law predicts a logarithmic distribution.  It doesn’t apply to all large datasets, but it does happen regularly enough that it is even used in detecting possible fraud in accounting, socio-economic data and elections.  In the United States, evidence based on Benford’s law is legally admissible in criminal cases at the federal, state, and local levels.

Testing Benford’s Law is a simple experiment to see how many large, publicly accessible datasets satisfy Benford’s Law.  Site creators, Jason Long and Bryce Thronton have tested a variety of large, publicly available datasets, including our Twitter Census: Twitter User by Friend Count dataset and found some striking results.