Using social engagement as a proxy indicator for consumer sentiment has often been a vaguely quantifiable task. That’s largely due to the inherently noisy nature of social content — sentiments are chaotic, spelling inconsistencies and typos are the norm, and unexpected user behavior seems to appear at random.
Large Numbers and Bespoke Filters
Stonewright’s technology uses multiple techniques to cut through the noise. A primary technique is to exploit Jakob Bernoulli’s 1713 Law of Large Numbers. In short, this Law tells us that large bodies of text will show a tendency to minimize the overall impact of random events. The Munchies Index is already over 42M words (as of November ’20) and it is expanding constantly. The larger the database becomes, the less affected it is by randomness.