Finding Hidden Corona Symptoms through Google Trends
A few days back, a study by Harvard was released which claimed that the infection had started back in August 2019. They used data for symptom-related queries on search engines for things such as “cough” and “diarrhoea” and satellite imagery of hospital parking lots in China’s Wuhan city — where the disease was first identified in late 2019.
This same methodology can be applied to Google Trends to check when symptom related queries spiked in different parts of the world. So, I decided to do this and publish the results.
Data & Findings
The query data for the last five years for India and USA has been used in this study.
As you can see, all the symptoms-related queries spike around March 2020. This also coincides with the spike in cases in US.
Next, let’s look at India’s data.
We can see is that in India, the results aren’t down yet. Hence, the first wave hasn’t ended.
Inferences from this data
Now, this data in itself isn’t great. We already know what symptoms of covid are. Right? but are there any symptoms which haven’t caught the expert’s eye? We can find them too using by examining trends for those symptoms.
For instance look at this graph which depicts the amount of searches for “back pain” done in US and India.
In India, this search lie well above the normal amount. But in US, there’s no increase in this search. Does this mean symptoms of corona changes with location? I don’t know. Maybe we’ll know in the coming months.
Practical Applications
- We can analyze different symptoms queries to see if they are related to covid.
- This data can be used by institutions to predict the coming wave and anticipate medical needs.
- Google Trends also has region filter. We can use it to monitor new outbreaks. If the search for symptoms goes up in any region, we can assume that that area is virus prone.
Thanks!
Mail me at prakhar897@gmail.com if you have any queries.
