Skip to main content
Fig. 1 | BMC Medical Research Methodology

Fig. 1

From: Need of care in interpreting Google Trends-based COVID-19 infodemiological study results: potential risk of false-positivity

Fig. 1

Outline of preprocessing flow. The sequence was processed to remove seasonality (1-year level) and general trend from the original series, and the remaining random series (A lowermost row) was used as the keyword trend data. Then, the obtained series were evaluated with an ADF test to examine its stationarity (B). Next, the temporal relationship between the processed sequence of each single keyword and the COVID-19 weekly positivity data was analyzed using the VAR model (C). Then we assessed whether the keyword trend Granger-causes COVID-19 positivity trends (D)

Back to article page