Skip to main content
Figure 2 | EPJ Data Science

Figure 2

From: Estimating suicide occurrence statistics using Google Trends

Figure 2

Estimating official suicide counts using Google search data. We investigate the relationship between the volume of searches for (A) ‘depression’ and (B) ‘suicide’ and the number of suicide occurrences, with a lag between the search data and the suicide data of up to 12 months (x-axis). Solid red lines represent the estimated effect and the dashed blue lines illustrate the 90% confidence intervals of these effects. At a lag of 0 months, we find that a higher number of searches for the term ‘depression’ corresponds to a lower number of suicides. However, this effect is not statistically significant. For lags of 5 to 10 months, a greater number of searches for the term ‘depression’ corresponds to a greater number of suicides. Conversely, at a lag of 0 months, we find that a greater number of searches for the term ‘suicide’ corresponds to a higher number of suicides. The effect vanishes after approximately 2 months. For lags of 6 to 11 months, a greater number of searches for the term ‘suicide’ corresponds to a lower number of suicides. However, it should be noted that the changes between positive and negative effects for both search terms may be due to the polynomial shape induced by the Almon model specification.

Back to article page