We study the impact of uncertainty shocks using a disaggregated model featuring US state-level unemployment, and uncertainty measured using state-level Google search data. Importantly, the model captures spillovers across states and identifies substantial differences in peak responses and time dynamics. Moreover, the importance of national factors in propagating the effect of uncertainty is also heterogeneous across states, but less relevant than state-level factors. These heterogeneous effects are related to state-specific industry compositions and fiscal positions. In addition, we highlight the usefulness of disaggregated data in models of uncertainty and economic activity for the USA, based on the superior ability of the disaggregate model to predict aggregate uncertainty and unemployment.

In nature conservation, the generation of public interest, attention or emotions is an important instrument for nature, biotope and species protection; in this, charismatic flagship species play an important role. In the present study, flagship-making affiliation to a taxonomic unit as well as morphological, ecological and conservation traits were identified by analysing vertebrate species from each of the five extant vertebrate classes (Mammalia, Aves, Reptilia, Amphibia and fishes). Google Trends data on the 20 most Googled species of each of the five classes were used, a representation index was derived and the body mass, diet and protection status of these species were analysed. A clear concentration of interest in mammalian species was evident with the help of the introduced representation index. Furthermore, species with a higher body mass were clearly overrepresented in the data. Overall, important patterns in the studied traits were determined: belonging to Mammalia, a large body mass and a carnivorous diet are frequently represented among these species. For conservation purposes, such popular species can be specifically selected as flagship species or ambassadors to help protect entire biomes, which will therefore benefit less charismatic species as well. Possible ways to use traits that are perceived to be flagship-making in order to further the global conservation endeavour are briefly discussed.

Every internet search query made out of curiosity by anyone who observed something in nature, as well as every photo uploaded to the internet, constitutes a data point of potential use to scientists. Researchers have now begun to exploit the vast online data accumulated through passive crowdsourcing for studies in ecology and epidemiology. Here, we demonstrate the usefulness of iParasitology, i.e. the use of internet data for tests of parasitological hypotheses, using hairworms (phylum Nematomorpha) as examples. These large worms are easily noticeable by people in general, and thus likely to generate interest on the internet. First, we show that internet search queries (collated with Google Trends) and photos uploaded to the internet (specifically, to the iNaturalist platform) point to parts of North America with many sightings of hairworms by the public, but few to no records in the scientific literature. Second, we demonstrate that internet searches predict seasonal peaks in hairworm abundance that accurately match scientific data. Finally, photos uploaded to the internet by non-scientists can provide reliable data on the host taxa that hairworms most frequently parasitize, and also identify hosts that appear to have been neglected by scientific studies. Our findings suggest that for any parasite group likely to be noticeable by non-scientists, information accumulating through internet search activity, photo uploads, social media or any other format available online, represents a valuable source of data that can complement traditional scientific data sources in parasitology.

Google search is ubiquitous, and Google Trends (GT) is a potentially useful access point for big data on many topics the world over. We propose a new ‘variance-in-time’ method for forecasting events using GT. By collecting multiple and overlapping samples of GT data over time, our algorithm leverages variation both in the mean and the variance of a search term in order to accommodate some idiosyncracies in the GT platform. To elucidate our approach, we use it to forecast protests in the United States. We use data from the Crowd Counting Consortium between 2017 and 2019 to build a sample of true protest events as well as a synthetic control group where no protests occurred. The model’s out-of-sample forecasts predict protests with higher accuracy than extant work using structural predictors, high frequency event data, or other sources of big data such as Twitter. Our results provide new insights into work specifically on political protests, while providing a general approach to GT that should be useful to researchers of many important, if rare, phenomena.

Zika virus infection in humans has been linked to severe neurological sequels and foetal malformations. The rapidly evolving epidemics and serious complications made the frequent updates of Zika virus mandatory. Web search query has emerged as a low-cost real-time surveillance system to anticipate infectious diseases’ outbreaks. Hence, we developed a prediction model that could predict Zika-confirmed cases based on Zika search volume in Google Trends. We extracted weekly confirmed Zika cases of two epidemic countries, Brazil and Colombia. We got the weekly Zika search volume in the two countries from Google Trends. We used standard time-series regression (TSR) to predict the weekly confirmed Zika cases based on the Zika search volume (Zika query). The basis TSR model – using 1-week lag of Zika query and using 1-week lag of Zika cases as a control for autocorrelation – was the best for predicting Zika cases in Brazil and Colombia because it balanced the performance of the model and the advance time in the prediction. Our results showed that we could use Google search queries to predict Zika cases 1 week earlier before the outbreak. These findings are important to help healthcare authorities evaluate the outbreak and take necessary precautions.

Explanatory models accounting for variation in policy choices by democratic governments usually include a demand (by the public) and a supply (by the government) component, whereas the latter component is usually better developed from a measurement viewpoint. The main reason is that public opinion surveys, the standard approach to measuring public demand, are expensive, difficult to implement simultaneously for different countries for purposes of crossnational comparison and impossible to implement ex post for purposes of longitudinal analysis if survey data for past time periods are lacking. We therefore propose a new approach to measuring public demand, focussing on political claims made by nongovernmental actors and expressed in the news. To demonstrate the feasibility and usefulness of our measure of published opinion, we focus on climate policy in the time period between 1995 and 2010. When comparing the new measure of published opinion with the best available public opinion survey and internet search data, it turns out that our data can serve as a meaningful proxy for public demand.