Ten years of research change using Google Trends: From the perspective of big data utilizations and applications

Jun, Seung-pyo; Yoo, Hyoung Sun; Choi, San

doi:10.1016/j.techfore.2017.11.009

Cited by 377 publications

(269 citation statements)

References 62 publications

Supporting

Mentioning

219

Contrasting

Unclassified

Order By: Relevance

“…If the data were retrieved each week during the season, results could have been different. [17] This is a point that should be assessed in the future.…”

Section: Discussionmentioning

confidence: 99%

PREPRINT: Using digital epidemiology methods to monitor influenza-like illness in the Netherlands in real-time: the 2017-2018 season

Schneider

Gool

Spreeuwenberg

et al. 2018

Preprint

View full text Add to dashboard Cite

IntroductionDespite the early development of Google Flu Trends in 2009, digital epidemiology methods have not been adopted widely, with most research focusing on the USA. In this article we demonstrate the prediction of real-time trends in influenza-like illness (ILI) in the Netherlands using search engine query data. MethodsWe used flu-related search query data from Google Trends in combination with traditional surveillance data from 40 general sentinel practices to build our predictive models. We introduced an artificial 4-week delay in the use of GP data in the models, in order to test the predictive performance of the search engine data.Simulating the weekly use of a prediction model across the 2017/2018 flu season we used lasso regression to fit 52 prediction models (one for each week) for weekly ILI incidence. We used rolling forecast cross-validation for lambda optimization in each model, minimizing the maximum absolute error. ResultsThe models accurately predicted the number of ILI cases during the 2017/18 ILI epidemic in real time with a mean absolute error of 1.40 (per 10,000 population) and a maximum absolute error of 6.36. The model would also have identified the onset, peak, and end of the epidemic with reasonable accuracyThe number of predictors that were retained in the prediction models was small, ranging from 3 to 5, with a single keyword ('Griep' = 'Flu') having by far the most weight in all models. DiscussionThis study demonstrates the feasibility of accurate real-time ILI incidence predictions in the Netherlands using internet search query data .Digital ILI monitoring strategies may be useful in countries with poor surveillance systems, or for monitoring emergent diseases, including influenza pandemics. We hope that this transparent and accessible case study inspires and supports further developments in field of digital epidemiology in Europe and beyond.

show abstract

“…If the data were retrieved each week during the season, results could have been different. [17] This is a point that should be assessed in the future.…”

Section: Discussionmentioning

confidence: 99%

PREPRINT: Using digital epidemiology methods to monitor influenza-like illness in the Netherlands in real-time: the 2017-2018 season

Schneider

Gool

Spreeuwenberg

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Here, it was demonstrated that the correspondence between SVI and real-world societal events can be explained by the fact that SVI tap into an online pool of shared ideas, values and representations (see Marzouki & Oullier, 2012;Marzouki & Oullier, 2014). Consequently, SVI analysis is now a well-established research tool for forecasting and monitoring social phenomena, with applications ranging from predicting disease contagion to modelling products 'hype cycles' in consumer research (for a review see Jun, Yoo, & Choi, 2018). SVI reflect fluctuating, aggregated search queries from individuals: they depict a part of what goes on in the mind of a given population at a given time.…”

Section: Svi Analysis In a Nutshellmentioning

confidence: 99%

“…SVI reflect fluctuating, aggregated search queries from individuals: they depict a part of what goes on in the mind of a given population at a given time. This is because search queries are behavioural indicators of epistemic motivations (information seeking; Jun et al, 2018). In that sense, SVI can be understood as probing a form of macro-social phenomenon emerging from individual aggregated behaviours.…”

Section: Svi Analysis In a Nutshellmentioning

confidence: 99%

“…When selecting this token we decided not to use it in combination with other tokens (e.g. TERROR+ATTACKS+ISIS), as sometimes done in the literature (Jun et al, 2018;Marzouki & Oullier, 2014). In fact, we needed the token to reflect general exposure to any threatening material related to terror attacks (and not necessarily those happening in France).…”

Section: Word Selectionmentioning

confidence: 99%

See 1 more Smart Citation

The dynamics of public opinion following terror attacks: Evidence for a decrease in equalitarian values from Internet Search Volume Indices

Troïan¹,

Arciszewski²,

Apostolidis³

2019

Cyberpsychology

View full text Add to dashboard Cite

The tendency of terror attacks to generate increased right-wing attitudes is a direct prediction from the motivated cognition approach to political ideology (Jost, 2017). However, due to methodological and theoretical problems, evidence for this 'right-shift' hypothesis is currently mixed. To address these issues, we introduce for the first-time search volume indices (SVI) analysis to the study of public opinion dynamics by assessing the effects of exposure to the 2015 Paris terror attacks, with a focus on equalitarian values related searches in the French cyberspace. Consistent with the 'right-shift' hypothesis, we demonstrate that online collective threat salience (SVI for the word 'terror attack') predicts significant decreases in equalitarian values SVI (e.g. 'equality') 6 weeks later, but not in nonequalitarian values SVI (e.g. 'liberty'). Mixed-model analyses of SVI for the period 2012-2017 confirmed these results by showing both decrease in equalitarian values SVI and no change in non-equalitarian values SVI after 2015. These findings corroborate the 'right-shift' hypothesis at the societal level using ecological-behavioural measures of public opinion and demonstrate the value of SVI analysis for theory-testing in political psychology. Data CollectionAccording to Google (2017), Trends data is 'an unbiased sample of Google search data. from both real-time random samples of the past week SVI and non-real-time data from a random sample of SVI 'pulled from as far back as 2004 and up to 36 hours before the search'. It is then 'adjusted proportionate to the time and location of a query by dividing the total searches of the geography and time range it represents to compare relative popularity […] The resulting numbers are then scaled on a range of 0 to 100 based on a topic's proportion to all searches on all topics.' Still according to the same source, Google Trends' algorithm excludes data from 'searches made by very few people […] duplicate searches […] repeated searches from the same person over a short period of time'. Time considerations. Google Trends SVI provides relative data: the weekly proportion of search index averaged by the total sum of search indexes, which changes every week. While this is not problematic for data analysis (and

show abstract

“…Additionally, it provides a new metric, changes in a web search index, to measure the impact of specific events on consumer discovery. Note that, while web search indexes have been already used in the literature mainly for forecasting purposes (Jun et al, 2018), to the best of our knowledge no paper has attempted to analyze the increase in public awareness, measured through search activity, induced by gatekeepers.…”

Section: Introductionmentioning

confidence: 99%

Music festivals as mediators and their influence on consumer awareness

Montoro–Pons

García

2020

Poetics

View full text Add to dashboard Cite

Cultural products compete for public awareness in markets with high uncertainty, oversupply and a short product life cycle. Altogether, this means that only a small fraction of all releases generate the necessary consumer awareness to achieve a significant commercial success. This paper aims at identifying the informational function music festivals serve in cultural markets and how it translates into consumer discovery of cultural supply. To do so, we empirically measure informational spillover effects to performers at an established music festival. We hypothesize that this effect stems from the reputation attached to the brand equity of cultural organizations, is asymmetric, as it decreases with the success of the performer such that lesser-known performers profit more than well-known ones. Empirical results are consistent with our hypotheses.

show abstract

Ten years of research change using Google Trends: From the perspective of big data utilizations and applications

Cited by 377 publications

References 62 publications

PREPRINT: Using digital epidemiology methods to monitor influenza-like illness in the Netherlands in real-time: the 2017-2018 season

PREPRINT: Using digital epidemiology methods to monitor influenza-like illness in the Netherlands in real-time: the 2017-2018 season

The dynamics of public opinion following terror attacks: Evidence for a decrease in equalitarian values from Internet Search Volume Indices

Music festivals as mediators and their influence on consumer awareness

Contact Info

Product

Resources

About