As regular COVID-19 testing winds down and hospital admissions continue to drop, experts are looking for new ways to track the virus. Some have even suggested that monitoring scented candle reviews might be useful for predicting future infection spikes.
It was a viral tweet that sparked the idea. A Portland science illustrator named Terri Nelson noticed that a number of angry customers were complaining that their Yankee Candles didn’t smell. She was inspired to look into the issue, and discovered that there seemed to be a correlation between the amount of bad smelling candles and the rise in reported COVID-19 cases.
Northeastern University political scientist Nick Beauchamp was a little skeptical at first, but he decided to dig into the data. Using a Chrome extension, he grabbed 9,800 reviews from Amazon for four best-selling Yankee candles and calculated the percentage of them that said “no scent.” He then plotted those results alongside reported weekly COVID-19 cases. His graph shows a clear spike around December, which coincides with the beginning of the Omicron wave.
He says that the trend seems to reflect anosmia, or loss of sense of smell, a symptom of the virus. The link was still present even after he controlled for the fact that more people buy and review candles in winter. Beauchamp says he’s not an epidemiologist, but that studying the data is interesting and could be helpful for other researchers. He also points out that examining Yankee Candle reviews is an easy way to spot trends because the company groups related comments together. He tells Protocol that he’s working on another project using alternative data sources to track COVID-19 trends.