A couple months ago, we discovered that Yahoo Mail users spend $110 more on electricity each year compared to Gmail users.
But some people inquired: what about looking “to the left of the @ sign”? How does that part of an email address correlate with energy usage?
So we examined correlations between the non-domain part of a household’s primary email address (e.g. “outlier” in firstname.lastname@example.org) and their electricity consumption.
For example, is having the word “sexy” in an email address associated with high electricity usage or low electricity usage?
For the purposes of comparison, we distinguished “high users” from “low users” by establishing a dividing line of 9,700 kWh/year, corresponding to the national average electricity consumption for homes in our email-domains dataset. Then we looked at which email-address keywords were correlated with usage above and below that level.
In the word clouds below, “sexy” is in the high-usage group (because email addresses containing the word “sexy” had average annual electricity usage of 10,201 kWh); “cute” is in the low-usage group (because email addresses containing the word “cute” had average annual usage of 9,146 kWh).
Some other intriguing discoveries…
· Email addresses with the word “coal” and “oil” are associated with annual high electricity use (11,953 kWh and 11,230 kWh), while “nuclear” is associated with lower use (8,295 kWh)
· “Hippies” (10,168 kWh) tend to be “high” users
· The average “enviro” is walking the walk, with the lowest annual usage of all names considered (8,054 kWh)
· Among mythical creatures, “unicorns” and “wizards” exhibit high usage, while “hobbits” appear to have lower electricity needs (perhaps because they have smaller abodes?)
Admittedly, these findings are based on a much smaller sample size (<50,000 data points) and are less statistically robust than the findings of our “Gmail versus Yahoo” analysis (>1.15 million data points).
But I am still convinced that a long-time hunch of mine has finally been validated: there is a correlation between being energy-efficient and being a “baller.”
Special thanks to Jillian Cairns and Efrat Levush.
Methodology: Prior to analysis, all email addresses were anonymized and separated from personally identifiable information. We selected fun names, but only included a name in the word clouds if there were at least 20 instances of it in our database (though for most names there were more than 200 instances). The email addresses considered here are not restricted to Yahoo Mail and Gmail. Larger words in the high-usage cloud are associated with the highest usage (e.g. “bbq”); larger words in the low-usage cloud are associated with the lowest usage (e.g. “enviro”). Note than annual average US household electricity consumption was 11,496 kWh as of 2010; we chose 9,700 kWh as a dividing “median” line because it reflects the average in our dataset, and assigned a similar number of items into each word cloud.
Data Privacy: All data analyzed here are anonymous and treated in strict adherence to Opower’s Data Principles.
Author’s note: The analysis and commentary presented above solely reflect the views of the author(s) and do not reflect the views of Opower’s utility partners.