Table 1 Results of model estimation.

From: An ecological and digital epidemiology analysis on the role of human behavior on the 2014 Chikungunya outbreak in Martinique

Parameters included in transmission rate Twitter as an anticipated indicator (τ = −1) Twitter as a real-time indicator (τ = 0) Twitter as a delayed indicator (τ = 1)
None 9088 9088 9088
Mosquito abundance (MA) 6980 6980 6980
Expressed protection need (EPN) 3897 7546 5191
Epidemics awareness (EA) 7797 8240 6878
MA and EPN 2402 4058 3161
MA and EA 7242 8518 5484
EPN and EA 7218 7529 3685
MA, EPN and EA 4389 7639 2675
  1. We show here the squared root of the Mean-Squared Error instead of AIC in order to show the difference between the observed and predicted number of cases. The best model includes the variation in mosquito abundance and the expressed need for protection represented by the proportion of tweets talking about protection against the mosquito in the set of all tweets and retweets that included the word Chikungunya (only Twitter accounts declared in Martinique have been considered). The lag period (τ) is expressed in month. AIC values are included in Supplementary Materials.