Google is releasing data on how people have been searching for COVID-19 symptoms, in the hope it will help researchers track the virus

• Google is releasing a huge dataset of search trends related to COVID-19 symptoms for researchers and public health authorities.
• The data, which Google promises is completely anonymized, will reveal trends for more than 400 symptoms.
• It hopes the data can be used to track how the virus is spreading.
• Are you a Google insider with more to share? Contact this reporter using encrypted email (hslangley@protonmail.com) or encrypted messaging apps Signal/Telegram (628-228-1836).

Google is releasing a database of US search trends for COVID-19 symptoms, hoping it will help public health authorities and researchers track how the virus is spreading.

Google built its dataset with user searches for more than 400 symptoms such as coughing, fever, and difficulty breathing. The aggregated data, which Google says is anonymized, shows trends in the volume of symptom-related searches at the US county level.

But Google says it isn't revealing the raw number of specific searches, and will instead normalize the search terms on a scale of 1 to 100, similar to how its Google Trends tool works, so researchers can identify spikes in search trends.

The data also won't show the specific searches themselves but will map them to symptoms and conditions that may relate to COVID-19.

The data will be made available through Google Cloud's dataset program as well as a GitHub repository. The dataset includes searches going back to 2017 so researchers can compare trends to previous years too.

But for now, the data is only for the US and includes data for searches made in English and Spanish. It will also only include search data for counties that meet a minimum search threshold, which Google says is designed to protect privacy. However, it did not say what that threshold was.

A spokesperson added that the data was produced using differential privacy – a technique of adding noise to the data to prevent it from being linked back to individual users.

"Researchers could use this dataset to study if search trends can provide an earlier and more accurate indication of the reemergence of the virus in different parts of the country," said Evgeniy Gabrilovich, a Google Health senior staff research scientist, in a statement.

Along with symptoms directly related to COVID-19, the data also includes searches for what Google is calling "secondary impacts" – conditions like anxiety and depression that have also risen during the pandemic.

This isn't the first time Google has leveraged its vast knowledge of user behavior for pandemic purposes. Earlier this year it started publishing "Community Mobility Reports," which use location data from users' devices to reveal how people are moving during COVID-19.

This also isn't the first time Google has tried to follow search trends related to diseases. A similar tool called Google Flu Trends, launched in 2008, attempted to track influenza trends based on search activity, but was shuttered in 2015 after it drastically overestimated the prevalence of flu over multiple years.