I collected data from TikTok for my logs project because I always scroll on the platform anyway. I watched 100 videos on my For You Page and noted the first number of likes in a chart in a Word document and compared them to the given Benford percents. Some discrepancies and reasons that his law may not properly reflect on the data gathered that I noticed include…

#1) Benford’s Law works well for data that is generated naturally, like populations, stock prices, or scientific measurements. Data that is bounded or constrained, such as prices of goods in a specific range, or data generated by human choice (e.g., TikTok likes), may not follow the expected distribution as closely. In datasets where human choice plays a significant role, Benford’s Law may be violated.

#2) Small datasets may not exhibit the expected first-digit distribution, especially if the data doesn’t span a wide range of values. In the video provided on teams, 5,150 pieces of data were collected compared to mine which was only 100.

Benford’s law can be used in fields such as…

#1) Fraud Detection: In finance and accounting, Benford’s Law can be used to detect anomalies in financial statements, tax returns, or other financial records. If the distribution of the first digits of reported figures deviates significantly from Benford’s Law, it could be an indicator of fraud or manipulation.

#2) Elections and Voting: Benford’s Law has been used to assess the authenticity of election data. If election results follow a distribution of first digits that diverges from Benford’s Law, it could suggest manipulation of the vote counts.

#3) Environmental Studies: Benford’s Law is sometimes used in analyzing data in the fields of geography. For example, in climate data, the distribution of 1st digits might reveal inconsistencies in data collection or reporting.

We also can use logs for Benford’s law. This captures the fact that smaller numbers like 1, 2, and 3 are more likely to appear as leading digits because they span a larger range of values on a scale using logs. The size of the difference between powers of 10 such as 1 to 10, or 10 to 100, grows exponentially, and logs naturally capture this scaling.

Here is my visual representation of the data I have collected.

print