Google Search Query Data Research Articles

Deep Learning Models for Airport Demand Forecasting With Google Trends

Managers gain new insights into how operational benefits can be achieved. Forecasting problems for passenger flow in airports are gaining interest among marketing researchers, but comparison of stochastic optimisation methods via deep learning forecasts with search query data is not yet available in the aviation field. To fill this gap, the current study predicts the demand of Madrid airport demand with Google search query data using H2O deep learning method. The findings indicate that there is a long-term relationship between search queries and actual passenger demand. Besides, search queries “fly to madrid,” and “flights to madrid spain” were found to be the cause of the actual domestic air passenger demand in Madrid. Also, to determine the best forecasting accuracy, stochastic gradient descent (SGD) optimisers were used. Specifically, findings indicate that Adam is a better optimiser increasing forecasting accuracy for Madrid airports.

Analysis of the Relationship Between COVID-19 Infections and Web-Based Housing Searches

Abstract The study used Google search query data on real estate interest for several countries in the Baltic area. The dynamics of public interest in housing have been compared to the dynamics of the COVID-19 infections in Lithuania, Latvia, Poland, and Sweden. This study uses the Vector autoregressive (VAR) model to forecast such time series. VAR is a multivariate linear time series model in which the endogenous variables in the system are lagged functions of the values of all endogenous variables. The increase in COVID-19 infections negatively affected society’s interest in housing. The study used Google Trends and R software.

Impact of the Food and Drug Administration enforcement policy on flavored e-cigarettes on the online popularity of disposable e-cigarettes: analyses of Google search query data

BackgroundThe impact of the U.S. Food and Drug Administration's (FDA) initial enforcement policy on flavored cartridge-based e-cigarettes and subsequent notice for the removal of flavored disposable electronic cigarettes (e-cigarettes) such as Puff Bar from the market has not been well evaluated. We, therefore, sought to examine the impact of the e-cigarette flavor-related policy changes on the online popularity of Puff Bar, a prototypic disposable e-cigarette.MethodsWe tabulated the total weekly Google search queries originating from the U.S. for "Puff Bar" and "Puff Bars" from January 1, 2019, to December 31, 2021. We divided the three years into four (4) distinct periods using the dates of the initial announcement to ban non-tobacco flavored e-cigarettes (September 11, 2019), the finalization of the FDA enforcement policy on cartridge-based flavors (January 2, 2020), and the notice for the market withdrawal of flavored disposable e-cigarettes (July 20, 2020) as reference time points. Then, we used piecewise linear regression and autoregressive integrated moving average (ARIMA) to compare the trends in searches for Puff Bar for the four (4) periods.ResultsBefore the initial announcement to ban non-tobacco flavored e-cigarettes, online search queries (per 10 million Google searches) for Puff Bar were slowly rising at a rate of 0.58 queries per week (95%CI: -0.80 – 1.97). Following the announcement, searches for Puff Bar increased significantly at a rate of 16.61 queries per week (95%CI: 12.13 – 21.10). The rate of searches for Puff Bar then increased exponentially at 40.08 queries per week (95%CI: 27.32 – 52.84) following the FDA flavor ban, which excluded disposable e-cigarettes. Then, the rate of increase declined but remained relatively stable at 3.67 queries per week (95%CI: 0.69–6.65) until the FDA's notice to remove flavored Puff Bar products from the market. Following this notice, the rate of searches for Puff Bar significantly declined (rate: -4.97 queries per week; 95%CI: -5.40—-4.54).ConclusionsThe tracking of online search data demonstrates rapid public recognition of the FDA's announcements of tobacco regulatory actions.

Using web search queries to monitor influenza-like illness: an exploratory retrospective analysis, Netherlands, 2017/18 influenza season.

BackgroundDespite the early development of Google Flu Trends in 2009, standards for digital epidemiology methods have not been established and research from European countries is scarce.AimIn this article, we study the use of web search queries to monitor influenza-like illness (ILI) rates in the Netherlands in real time.MethodsIn this retrospective analysis, we simulated the weekly use of a prediction model for estimating the then-current ILI incidence across the 2017/18 influenza season solely based on Google search query data. We used weekly ILI data as reported to The European Surveillance System (TESSY) each week, and we removed the then-last 4 weeks from our dataset. We then fitted a prediction model based on the then-most-recent search query data from Google Trends to fill the 4-week gap (‘Nowcasting’). Lasso regression, in combination with cross-validation, was applied to select predictors and to fit the 52 models, one for each week of the season.ResultsThe models provided accurate predictions with a mean and maximum absolute error of 1.40 (95% confidence interval: 1.09–1.75) and 6.36 per 10,000 population. The onset, peak and end of the epidemic were predicted with an error of 1, 3 and 2 weeks, respectively. The number of search terms retained as predictors ranged from three to five, with one keyword, ‘griep’ (‘flu’), having the most weight in all models.DiscussionThis study demonstrates the feasibility of accurate, real-time ILI incidence predictions in the Netherlands using Google search query data.

Can One Improve Now-Casts of Crop Prices in Africa? Google Can.

With increasing Internet user rates across Africa, there is considerable interest in exploring new, online data sources. Particularly, search engine metadata, i.e. data representing the contemporaneous online-interest in a specific topic, has gained considerable interest, due to its potential to extract a near real-time online signal about the current interest of a society. The objective of this study is to analyze whether search engine metadata in the form of Google Search Query (GSQ) data can be used to improve now-casts of maize prices in nine African countries, these are Ethiopia, Kenya, Malawi, Mozambique, Rwanda, Tanzania and Uganda, Zambia and Zimbabwe. We formulate as benchmark an auto-regressive model for each country, which we subsequently augment by two specifications based on contemporary GSQ data. We test the models in in-sample, and in a pseudo out-of-sample, one-step-ahead now-casting environment and compare their forecasting errors. The GSQ specifications improve the now-casting fit in 8 out 9 countries and reduce the now-casting error between 3% and 23%. The largest improvement of maize price now-casts is achieved for Malawi, Kenya, Zambia and Tanzania, with improvements larger than 14%.

Accurate quantification of uncertainty in epidemic parameter estimates and predictions using stochastic compartmental models.

Stochastic transmission dynamic models are needed to quantify the uncertainty in estimates and predictions during outbreaks of infectious diseases. We previously developed a calibration method for stochastic epidemic compartmental models, called Multiple Shooting for Stochastic Systems (MSS), and demonstrated its competitive performance against a number of existing state-of-the-art calibration methods. The existing MSS method, however, lacks a mechanism against filter degeneracy, a phenomenon that results in parameter posterior distributions that are weighted heavily around a single value. As such, when filter degeneracy occurs, the posterior distributions of parameter estimates will not yield reliable credible or prediction intervals for parameter estimates and predictions. In this work, we extend the MSS method by evaluating and incorporating two resampling techniques to detect and resolve filter degeneracy. Using simulation experiments, we demonstrate that an extended MSS method produces credible and prediction intervals with desired coverage in estimating key epidemic parameters (e.g. mean duration of infectiousness and R0) and short- and long-term predictions (e.g. one and three-week forecasts, timing and number of cases at the epidemic peak, and final epidemic size). Applying the extended MSS approach to a humidity-based stochastic compartmental influenza model, we were able to accurately predict influenza-like illness activity reported by U.S. Centers for Disease Control and Prevention from 10 regions as well as city-level influenza activity using real-time, city-specific Google search query data from 119 U.S. cities between 2003 and 2014.

How Unemployment Affects Bond Prices: A Mixed Frequency Google Nowcasting Approach

We analyze the relationship between unemployment rate changes and government bond yields during and after the most recent financial crisis across ten industrialized countries. The study is conducted on a weekly basis and we therefore nowcast unemployment data, which are only available once a month, on a weekly frequency using Google search query data. In order to account for the time series’ long-memory components during the first-stage nowcasting and the second-stage modeling, we draw on Corsi’s (J Financ Economet 7:174–196, 2009) heterogeneous autoregressive time series model. In particular, we adapt this idea to a setting of mixed-frequency nowcasting. Our results indicate that Google searches greatly increase the nowcasting accuracy of unemployment rate changes. The impact of an idiosyncratic rise in unemployment on bond yields turns out to be positive for European countries while it is negative for the United States and Australia. The speed of the response also varies. Not unexpectedly, bond yields do not have an impact on unemployment. Our findings have interesting implications for the way shocks are absorbed in economic systems that differ, in particular, with respect to the central bank’s core tasks.

Google Trends and the forecasting performance of exchange rate models

AbstractIn this paper, we use Google Trends data for exchange rate forecasting in the context of a broad literature review that ties the exchange rate movements with macroeconomic fundamentals. The sample covers 11 OECD countries’ exchange rates for the period from January 2004 to June 2014. In out‐of‐sample forecasting of monthly returns on exchange rates, our findings indicate that the Google Trends search query data do a better job than the structural models in predicting the true direction of changes in nominal exchange rates. We also observed that Google Trends‐based forecasts are better at picking up the direction of the changes in the monthly nominal exchange rates after the Great Recession era (2008–2009). Based on the Clark and West inference procedure of equal predictive accuracy testing, we found that the relative performance of Google Trends‐based exchange rate predictions against the null of a random walk model is no worse than the purchasing power parity model. On the other hand, although the monetary model fundamentals could beat the random walk null only in one out of 11 currency pairs, with Google Trends predictors we found evidence of better performance for five currency pairs. We believe that these findings necessitate further research in this area to investigate the extravalue one can get from Google search query data.

Using Deep Learning Techniques to Forecast Environmental Consumption Level

Artificial intelligence is a promising futuristic concept in the field of science and technology, and is widely used in new industries. The deep-learning technology leads to performance enhancement and generalization of artificial intelligence technology. The global leader in the field of information technology has declared its intention to utilize the deep-learning technology to solve environmental problems such as climate change, but few environmental applications have so far been developed. This study uses deep-learning technologies in the environmental field to predict the status of pro-environmental consumption. We predicted the pro-environmental consumption index based on Google search query data, using a recurrent neural network (RNN) model. To verify the accuracy of the index, we compared the prediction accuracy of the RNN model with that of the ordinary least square and artificial neural network models. The RNN model predicts the pro-environmental consumption index better than any other model. We expect the RNN model to perform still better in a big data environment because the deep-learning technologies would be increasingly sophisticated as the volume of data grows. Moreover, the framework of this study could be useful in environmental forecasting to prevent damage caused by climate change.

They're heating up: Internet search query trends reveal significant public interest in heat-not-burn tobacco products.

Heat-not-burn tobacco products, battery powered devices that heat leaf tobacco to approximately 500 degrees Fahrenheit to produce an inhalable aerosol, are being introduced in markets around the world. Japan, where manufacturers have marketed several heat-not-burn brands since 2014, has been the focal national test market, with the intention of developing global marketing strategies. We used Google search query data to estimate, for the first time, the scale and growth potential of heat-not-burn tobacco products. Average monthly searches for heat-not-burn products rose 1,426% (95%CI: 746,3574) between their first (2015) and second (2016) complete years on the market and an additional 100% (95%CI: 60, 173) between the products second (2016) and third years on the market (Jan-Sep 2017). There are now between 5.9 and 7.5 million heat-not-burn related Google searches in Japan each month based on September 2017 estimates. Moreover, forecasts relying on the historical trends suggest heat-not-burn searches will increase an additional 32% (95%CI: -4 to 79) during 2018, compared to current estimates for 2017 (Jan-Sep), with continued growth thereafter expected. Contrasting heat-not-burn’s rise in Japan to electronic cigarettes’ rise in the United States we find searches for heat-not-burn eclipsed electronic cigarette searches during April 2016. Moreover, the change in average monthly queries for heat-not-burn in Japan between 2015 and 2017 was 399 (95% CI: 184, 1490) times larger than the change in average monthly queries for electronic cigarettes in the Unites States over the same time period, increasing by 2,956% (95% CI: 1729, 7304) compared to only 7% (95% CI: 3,13). Our findings are a clarion call for tobacco control leaders to ready themselves as heat-not-burn tobacco products will likely garner substantial interest as they are introduced into new markets. Public health practitioners should expand heat-not-burn tobacco product surveillance, adjust existing tobacco control strategies to account for heat-not-burn tobacco products, and preemptively study the health risks/benefits, popular perceptions, and health messaging around heat-not-burn tobacco products.

Using Twitter sentiment and emotions analysis of Google Trends for decisions making

PurposeAn ever-growing body of knowledge demonstrates the correlation among real-world phenomena and search query data issued on Google, as showed in the literature survey introduced in the following. The purpose of this paper is to introduce a pipeline, implemented as a web service, which, starting with recent Google Trends, allows a decision maker to monitor Twitter’s sentiment regarding these trends, enabling users to choose geographic areas for their monitors. In addition to the positive/negative sentiments about Google Trends, the pipeline offers the ability to view, on the same dashboard, the emotions that Google Trends triggers in the Twitter population. Such a set of tools, allows, as a whole, monitoring real-time on Twitter the feelings about Google Trends that would otherwise only fall into search statistics, even if useful. As a whole, the pipeline has no claim of prediction over the trends it tracks. Instead, it aims to provide a user with guidance about Google Trends, which, as the scientific literature demonstrates, is related to many real-world phenomena (e.g. epidemiology, economy, political science).Design/methodology/approachThe proposed experimental framework allows the integration of Google search query data and Twitter social data. As new trends emerge in Google searches, the pipeline interrogates Twitter to track, also geographically, the feelings and emotions of Twitter users about new trends. The core of the pipeline is represented by a sentiment analysis framework that make use of a Bayesian machine learning device exploiting deep natural language processing modules to assign emotions and sentiment orientations to a collection of tweets geolocalized on the microblogging platform. The pipeline is accessible as a web service for any user authorized with credentials.FindingsThe employment of the pipeline for three different monitoring task (i.e. consumer electronics, healthcare, and politics) shows the plausibility of the proposed approach in order to measure social media sentiments and emotions concerning the trends emerged on Google searches.Originality/valueThe proposed approach aims to bridge the gap among Google search query data and sentiments that emerge on Twitter about these trends.

THE FORECASTING POWER OF INTERNET SEARCH QUERIES IN THE BRAZILIAN FINANCIAL MARKET

ABSTRACT Purpose: To analyze the predictability of Google's search queries in the Brazilian financial market. Originality/gap/relevance/implications: Despite a growing foreign literature using Google's search query data, there is no acknowledgement of work on this area in Brazil. An application to the Brazilian financial market shows new sources of information about market movements and may contribute to researchers and practitioners to understand how changes in specific search queries affect the market. Key methodological aspects: Following previous studies, we estimate VAR models and Granger causality tests to investigate the effects over three variables in both stock and fixed income markets: traded volume, return and volatility. Following this procedure, we verify both the hypothesis of financial variables being affected by search queries, as well as the opposite relationship. Weekly data from Google's search queries and financial markets was gathered for the period between 2007 and 2014. Summary of key results: The existence of a predictive effect between search query data and financial variables, particularly in the stock market, is evident. However, this result was not robust in all cases studied. It is noteworthy that, for the inverse relationship, i.e. financial market impacting search queries on Google, strong evidence of a causal relationship has been found. A trading strategy based on this type of data yielded higher returns than the defined benchmarks. Key considerations/conclusions: A significant relationship between Google's search query data and the financial market has been discovered. Results provide a new source of information that affects the Brazilian financial market.

An analysis of morbidity and mortality using Google Trends

ABSTRACTImprovement in information technology has led to an increase in the ability of individuals to access health information; however, there still exists a large gap between patient and physician knowledge about diseases and/or illnesses. Often patients find themselves in doubt or overwhelmed by the care they are given. High medical costs also persist which makes the experience of patients less than ideal. For these reasons, among others, the internet has become a very popular tool in the hands of individuals who may wish to better understand the status of their health, make more informed decisions about their treatment options, and/or find cheaper ways to maintain good health status or cure a disease. Internet search engines, especially Google, play a crucial role in providing a means for patients to get the health information they need. The analysis presented in this article seeks to investigate the possibility of using Google Trends data to analyze the relationship between internet searches on the major causes of death and actual mortality in the United States. Specifically, this research seeks to investigate how well Google search query data serve as proxy for data on the actual causes of death.

Mortgage default risk: New evidence from internet search queries

We use Google search query data to develop a broad-based and real-time index of mortgage default risk. Unlike established indicators, our Mortgage Default Risk Index (MDRI) directly reflects households’concerns regarding their risk of mortgage default. The MDRI predicts housing returns, mortgage delinquency indicators, and subprime credit default swaps. These results persist both in- and out-of-sample and at multiple data frequencies. Together, research findings suggest internet search queries yield valuable new insights into household mortgage default risk.

Sentiment-based predictions of housing market turning points with Google trends

Purpose– Recent research has found significant relationships between internet search volume and real estate markets. This paper aims to examine whether Google search volume data can serve as a leading sentiment indicator and are able to predict turning points in the US housing market. One of the main objectives is to find a model based on internet search interest that generates reliable real-time forecasts.Design/methodology/approach– Starting from seven individual real-estate-related Google search volume indices, a multivariate probit model is derived by following a selection procedure. The best model is then tested for its in- and out-of-sample forecasting ability.Findings– The results show that the model predicts the direction of monthly price changes correctly, with over 89 per cent in-sample and just above 88 per cent in one to four-month out-of-sample forecasts. The out-of-sample tests demonstrate that although the Google model is not always accurate in terms of timing, the signals are always correct when it comes to foreseeing an upcoming turning point. Thus, as signals are generated up to six months early, it functions as a satisfactory and timely indicator of future house price changes.Practical implications– The results suggest that Google data can serve as an early market indicator and that the application of this data set in binary forecasting models can produce useful predictions of changes in upward and downward movements of US house prices, as measured by the Case–Shiller 20-City House Price Index. This implies that real estate forecasters, economists and policymakers should consider incorporating this free and very current data set into their market forecasts or when performing plausibility checks for future investment decisions.Originality/value– This is the first paper to apply Google search query data as a sentiment indicator in binary forecasting models to predict turning points in the housing market.

Examining the Relationship Between Past Orientation and US Suicide Rates: An Analysis Using Big Data-Driven Google Search Queries.

BackgroundInternet search query data reflect the attitudes of the users, using which we can measure the past orientation to commit suicide. Examinations of past orientation often highlight certain predispositions of attitude, many of which can be suicide risk factors.ObjectiveTo investigate the relationship between past orientation and suicide rate by examining Google search queries.MethodsWe measured the past orientation using Google search query data by comparing the search volumes of the past year and those of the future year, across the 50 US states and the District of Columbia during the period from 2004 to 2012. We constructed a panel dataset with independent variables as control variables; we then undertook an analysis using multiple ordinary least squares regression and methods that leverage the Akaike information criterion and the Bayesian information criterion.ResultsIt was found that past orientation had a positive relationship with the suicide rate (P≤.001) and that it improves the goodness-of-fit of the model regarding the suicide rate. Unemployment rate (P≤.001 in Models 3 and 4), Gini coefficient (P≤.001), and population growth rate (P≤.001) had a positive relationship with the suicide rate, whereas the gross state product (P≤.001) showed a negative relationship with the suicide rate.ConclusionsWe empirically identified the positive relationship between the suicide rate and past orientation, which was measured by big data-driven Google search query.

Adoption of green electricity policies: Investigating the role of environmental attitudes via big data-driven search-queries

Despite the rising influence of public opinion on government energy policy formulation and implementation, the roles of pro and/or anti-environmental attitudes among residents have not been empirically examined. To quantify time-varying environmental attitudes among local residents, we exploit geo-specific Google search-query data derived from Internet-based “big data” and verify through ordinary least squares regression outcomes regarding environmental behavior. For the purpose of drawing policy implications, we revisit decisions by state governments of the United States to adopt three well-known green electricity policies: renewable energy portfolio, net metering rules, and public benefit funds. As some states have not yet adopted some (or any) of these policies, unlike previous studies, we handle the issue by examining right-censored data and applying a duration-based econometric method called the accelerated failure time model. We found state residents’ environmental attitudes to have statistically significant roles, after controlling for other traditional time-varying policy adoption factors. Interestingly, the extent to which anti-environmental attitudes affect a state’s policy adoption differs across green energy policies, and knowing this can help a local government formulate better-tailored environmental policy. In particular, researchers can use our method of incorporating citizens’ environmental attitudes to discuss relevant issues in the field of energy policy.

Nowcasting Mobile Games Ranking Using Web Search Query Data

In recent years, the Internet has become embedded into the purchasing decision of consumers. The purpose of this paper is to study whether the Internet behavior of users correlates with their actual behavior in computer games market. Rather than proposing the most accurate model for computer game sales, we aim to investigate to what extent web search query data can be exploited to nowcast (contraction of “now” and “forecasting” referring to techniques used to make short-term forecasts) (predict the present status of) the ranking of mobile games in the world. Google search query data is used for this purpose, since this data can provide a real-time view on the topics of interest. Various statistical techniques are used to show the effectiveness of using web search query data to nowcast mobile games ranking.

Mortgage Default Risk: New Evidence from Internet Search Queries

We use Google search query data to develop a broad-based and real-time index of mortgage default risk. Unlike established indicators, our Mortgage Default Risk Index (MDRI) directly reflects households' concerns regarding their risk of mortgage default. The MDRI predicts housing returns, mortgage delinquency indicators, and subprime credit default swaps. These results persist both in- and out-of-sample and at multiple data frequencies. Together, research findings suggest internet search queries yield valuable new insights into household mortgage default risk.

A Cross-Country Analysis of Unemployment and Bonds with Long-Memory Relations

We analyze the relationship between unemployment rate changes and government bond yields during and after the most recent financial crisis across nine industrialized countries. The study is conducted on a weekly basis and we therefore nowcast unemployment data, which are only available once a month, on a weekly frequency using Google search query data. In order to account for the time series' long-memory components during the first-stage nowcasting and the second-stage modeling, we draw on Corsi's (2009) heterogeneous autoregressive time series model. In particular, we adapt this idea to a setting of mixed-frequency nowcasting. Our results indicate that Google searches greatly increase the nowcasting accuracy of unemployment rate changes. The impact of an idiosyncratic rise in unemployment on bond yields turns out to be positive for European countries while it is negative for the United States and Australia. The speed of the response also varies. Not unexpectedly, bond yields do not have an impact on unemployment. Our findings have interesting implications for the way shocks are absorbed in economic systems that differ, in particular, with respect to the central bank's core tasks.