Research on Mercury Retrograde by Renay Oshop
Major Update in this Important Research!
Building a 99.6% Accurate Peak Date Detector for Misspellings in Amazon Reviews Using Astrology
Abstract: Almost 19 years of data were released by the online store Amazon. We paid attention particularly to the text reviews. Focusing on a set of 100 most common misspelling mistakes, we looked at the rate of misspellings (words misspelled divided by total words in the review) for each entry, averaged that across all entries for a day, and built a series across time.
Eleven yogas were found to occur with a relatively small p-value (three were less than 0.05) on the peak misspelling days. A predictor engine was built using the random forest method based on these yogas. The resultant artificial intelligence was able to predict peak dates and non-peak dates with 99.6 percent accuracy in cross validation.
We believe that this exercise in signal processing, artificial intelligence, and big data can show others an exciting new way for doing astrological research. The good results also can move the needle in contemporary discourse on the utility of astrology
“Mercury Retrograde;” when the planet Mercury appears to slow to a stop and start moving backwards, is a time of great concern in both western and Indian/Vedic astrological cultures.
Of course, everyone knows that the planet Mercury isn't actually going backwards. The apparent retrograde motion is one of many geocentric optical illusions we earthlings experience on a predictably regular basis.
To test the general thesis of an effect on communications during the Mercury Retrograde period, Renay Oshop analyzed a randomized 3.13836% of the Reddit comments in May 2015, wherein 31,863,786 submissions (approx. 1.8 million per day) occurred during the non-Mercury Retrograde period (aka Direct), while a further 22,640,624 (approx. 2.1 million per day) occurred when Mercury was retrograde. After exhaustive calculations (see below), Renay found a statistically significant difference (up to 2.14%) in the incidence of word spelling errors (divided by entry word length) between Mercury Retrograde (MR) and Non-Mercury Retrograde (NMR) entries.
To further test the general thesis of an effect on communications during Mercury Retrograde periods, Renay also analyzed a randomized 5% of a collection of 69,242,585 Amazon reviews posted between Jan 1, 2009 and July 23, 2014. (Amazon.com: Online Shopping for Electronics, Apparel, Computers, Books, DVDs & more).
Of course, the results of these two studies don't prove that Mercury retrograde causes spelling errors. Still, an important first step has been established. We now have some evidence that the period known as 'Mercury Retrograde' coincides with an increase in spelling errors and the misuse of non-dictionary words.
These results should be of great interest to both western and Vedic astrologers; corroborating and validating an essential element of their practice. Further, if this small, but significant increase in error rates were found to hold across a broader series of actions; potentially compounding throughout the number of small things we all do each day, it could increase the likelihood of an eventual big mistake during the 21 - 23 days of the Mercury Retrograde period. The answer lies in further testing and replication.
Below you will find more information on Renay Oshop, her background and qualifications, and links to the complete studies with all the relevant methods, materials and further discussions.
was accepted to MIT, Oxford, and Harvard at age 15 and accepted a faculty position at the University of Colorado Health Sciences Center at age 21 to specialize in early bioinformatics.
She subsequently attended Dr. Vasant Lad's Ayurvedic Institute, a kind of Hogswart's, and embarked on an international 18 years-and-counting journey to study Jyotisha