Figure 7
Block diagram, identifying the steps of the workflow. The number of sites to be included in the initial analysis was arbitrarily set in SEMrush to 2000. Senticnet extracted the count for each pool of words classified as pertaining to one emotion. We plotted the relative occurrence of pooled emotion-related words (columns) with each site (rows) for each condition (one spreadsheet for condition) in Excel, for English sites. The fingerprint graphic format was developed for a qualitative representation of each condition emotional pattern. The variables were confronted on a statistical level (Mann–Whitney U-test) for association significance and were forwarded to SVM linear testing.
