How machine learning addressed an ancient claim

Descriptions of 7,152 mundane events correlate with celestial positions.

by Renay Oshop

Correlation, Vol. 34(2) 2022, pp 35-41


The belief in a correspondence between the planetary positions at the time of a world event and those at a similar event at another time is ancient and still claimed today (Tarnas, 2006). Using a database of 7,152 dates of world events from Wolfram Research, the short descriptions that accompany each date are quantified using a technique known as embeddings. Embeddings, a gift from machine learning, are a way to measure similarities between text. In addition, the celestial latitudes of the Sun, Moon, planets, Pluto, and the Lunar Nodes are calculated for midnight on the start dates for each event. Significance is found in an appropriate test of independence (Spearman’s rank, p < .001; rho = .082) of the descriptive distance to the collective mean for each world event and the chart distance to the collective mean of the celestial positions on the day of the event.