Search Engine Data Research Articles

Interpretable tourism volume forecasting with multivariate time series under the impact of COVID-19.

This study proposes a novel interpretable framework to forecast the daily tourism volume of Jiuzhaigou Valley, Huangshan Mountain, and Siguniang Mountain in China under the impact of COVID-19 by using multivariate time-series data, particularly historical tourism volume data, COVID-19 data, the Baidu index, and weather data. For the first time, epidemic-related search engine data is introduced for tourism demand forecasting. A new method named the composition leading search index-variational mode decomposition is proposed to process search engine data. Meanwhile, to overcome the problem of insufficient interpretability of existing tourism demand forecasting, a new model of DE-TFT interpretable tourism demand forecasting is proposed in this study, in which the hyperparameters of temporal fusion transformers (TFT) are optimized intelligently and efficiently based on the differential evolution algorithm. TFT is an attention-based deep learning model that combines high-performance forecasting with interpretable analysis of temporal dynamics, displaying excellent performance in forecasting research. The TFT model produces an interpretable tourism demand forecast output, including the importance ranking of different input variables and attention analysis at different time steps. Besides, the validity of the proposed forecasting framework is verified based on three cases. Interpretable experimental results show that the epidemic-related search engine data can well reflect the concerns of tourists about tourism during the COVID-19 epidemic.

How to capture tourists’ search behavior in tourism forecasts? A two-stage feature selection approach

Search engine data have been widely used and shown to be useful in tourism demand forecasting. However, considering of the vast amounts of search keywords, how to better capture the tourists’ attention and explore the most predictive keyword combination remain unsolved. In this study, a two-stage feature selection-based methodology is proposed to address this question. Specifically, i.e., single feature selection method comparison for selecting a relative effective way to reduce the data dimension and ensure the quality of the initial subset, genetic algorithm in the second stage for obtaining feature subset better suitable for forecasting model with stronger predictive power. Experimental results indicate that the two-stage feature selection method outperforms all the considered benchmarks.

Using Google Trends Data to Learn More About Survey Participation

As response rates continue to decline, the need to learn more about the survey participation process remains an important task for survey researchers. Search engine data may be one possible source for learning about what information some potential respondents are looking up about a survey when they are making a participation decision. In the present study, we explored the potential of search engine data for learning about survey participation and how it can inform survey design decisions. We drew on freely available Google Trends (GT) data to learn about the use of Google Search with respect to our case study: participation in the Family Research and Demographic Analysis (FReDA) panel survey. Our results showed that some potential respondents were using Google Search to gather information on the FReDA survey. We also showed that the additional data obtained via GT can help survey researchers to discover topics of interest to respondents and geographically stratified search patterns. Moreover, we introduced different approaches for obtaining data via GT, discussed the challenges that come with these data, and closed with practical recommendations on how survey researchers might utilize GT data to learn about survey participation.

The Role of Information Boxes in Search Engine Results for Symptom Searches: Analysis of Archival Data.

Search engines provide health information boxes as part of search results to address information gaps and misinformation for commonly searched symptoms. Few prior studies have sought to understand how individuals who are seeking information about health symptoms navigate different types of page elements on search engine results pages, including health information boxes. Using real-world search engine data, this study sought to investigate how users searching for common health-related symptoms with Bing interacted with health information boxes (info boxes) and other page elements. A sample of searches (N=28,552 unique searches) was compiled for the 17 most common medical symptoms queried on Microsoft Bing by users in the United States between September and November 2019. The association between the page elements that users saw, their characteristics, and the time spent on elements or clicks was investigated using linear and logistic regression. The number of searches ranged by symptom type from 55 searches for cramps to 7459 searches for anxiety. Users searching for common health-related symptoms saw pages with standard web results (n=24,034, 84%), itemized web results (n=23,354, 82%), ads (n=13,171, 46%), and info boxes (n=18,215, 64%). Users spent on average 22 (SD 26) seconds on the search engine results page. Users who saw all page elements spent 25% (7.1 s) of their time on the info box, 23% (6.1 s) on standard web results, 20% (5.7 s) on ads, and 10% (10 s) on itemized web results, with significantly more time on the info box compared to other elements and the least amount of time on itemized web results. Info box characteristics such as reading ease and appearance of related conditions were associated with longer time on the info box. Although none of the info box characteristics were associated with clicks on standard web results, info box characteristics such as reading ease and related searches were negatively correlated with clicks on ads. Info boxes were attended most by users compared with other page elements, and their characteristics may influence future web searching. Future studies are needed that further explore the utility of info boxes and their influence on real-world health-seeking behaviors.

Forecast daily tourist volumes during the epidemic period using COVID-19 data, search engine data and weather data

The COVID-19 epidemic has brought a devastating blow to the tourism industry. Affected by the epidemic situation, the change of tourism volume of scenic spots is very unstable. Therefore, forecasting tourist volume in the context of COVID-19 epidemic is a new and challenging problem. In response, a novel multivariate time series forecasting framework based on variational mode decomposition (VMD) and gated recurrent unit network (GRU), i.e., VMD-GRU, is proposed to forecast daily tourist volumes during the epidemic. It takes the lead in using COVID-19 data, search traffic data and weather data. Through sufficient experiments and comparisons, the superiority of the approach is illustrated, and the predictive power of the above three types of data, especially the COVID-19 data, is revealed. Accurate forecast results from the method can help relevant government officials and tourism practitioners to better adjust tourism resources, cooperate with anti-epidemic work and reduce operational risks.

Intimate Partner Violence as Reflected in Internet Search Data

Intimate partner violence (IPV) is a major public health concern with serious consequences for victims’ physical and mental health. Despite the high prevalence of IPV, describing it and detecting people suffering from it is difficult due to its sensitive nature and stigma associated with it. Existing tools for screening and tracking IPV victims are laborious, time-consuming, expensive, and require human supervision. Search engine data has previously been shown to provide insights into temporal behavioral information about millions of users who are truthful in their information seeking. Here, we present a large-scale analysis of individuals experiencing IPV providing insights into their characteristics and behavior. We extracted the queries from Bing search engine data of more than 50 thousand US-based individuals suffering from IPV. We provide insights into the differences among subpopulations of people experiencing IPV and the topics they search for. We find that approximately half of the users begin to search for IPV following an acute event (physical violence or abuse) and 20% of users actively hide their interest in IPV. Users typically begin to seek help 3 weeks after beginning to query about IPV. The topics of interest to people experience IPV include the effects of IPV, help seeking, and methods to escape from IPV. Our insights show that early cues of IPV may be difficult to detect within search queries, though even in the late stage that many IPV users are identified, interventions such as ads to guide people to safely exit violent situations could be beneficial.

Uncertainty, Search Engine Data, and Stock Market Returns During a Pandemic.

In recent years, a series of uncertain events, including the spread of COVID-19, has affected the Chinese stock market. When people face uncertainty, they often turn to internet search engines to obtain more information to support their investment decisions. This paper uses the uncertainty index, investor sentiment reflected by search engine data, and Chinese stock return data during the pandemic to examine the relationships among the three. Using daily data from March 2, 2020, to March 2, 2021, our empirical findings reveal that stock returns during a pandemic lead to an increase in investor retrieval of search engine data and that uncertainty affects stock returns during a pandemic. However, the reverse is not true. Therefore, in the face of an uncertainty such as market volatility caused by the spread of the pandemic, the active release of favorable information by regulators can help guide investor sentiment, prevent sharp stock market volatility, and improve the effectiveness of policy governance.

Comparisons and predictions of intercity population migration propensity in major urban clusters in China: based on use of the Baidu index

The development of the Internet and big data have made it possible to study population migration and flow between cities. This study analyzes the probability of the population migration propensity of China’s three major urban clusters, identifies the direction of population movements, and uses Markov chains to predict the probability of population migration propensity moving forward in order to assess the intercity population migration trends of urban clusters in the future. Internet search engine data is used, and a population migration propensity intensity model is used for calculations. The results show that the Pearl River Delta urban cluster and the Yangtze River Delta urban clusters are areas of active population migration, and that intercity population migration is a part of this activity. Intercity population migration in the Beijing-Tianjin-Hebei urban cluster is not as active as it is in the Yangtze River Delta and the Pearl River Delta urban cluster. Although the physical distance between Beijing and surrounding cities is relatively small, the correlation degree of migration propensity is not high. In the future, Shanghai, Nanjing, and Zhoushan in the Yangtze River Delta urban cluster; Zhuhai, Shenzhen, Guangzhou, Huizhou, and Zhongshan in the Pearl River Delta urban cluster; and Tianjin in the Beijing-Tianjin-Hebei urban cluster will be the main destinations of China’s population migration.

The national burden of scabies in Germany: a population-based approach using Internet search engine data

PurposeScabies is a World Health Organization-defined neglected tropical disease and a growing public health issue worldwide. It is difficult to obtain reliable data on prevalence due to the lack of standardized tests. The aim of this study was to assess scabies online search behavior in Germany to identify local differences using Google search volume.MethodsGoogle Ads Keyword Planner was used to investigate the scabies-related search volume for Germany as a whole, its 16 federal states, and 15 large cities for the period from January 2016 to December 2019. The identified search terms were qualitatively categorized and critically analyzed.ResultsA total of 572 keywords with an overall search volume of 11,414,180 searches regarding scabies were identified in Germany. The number of searches was higher in winter than in summer, with a national peak in March 2018. Around 30.6% of the searches regarding scabies therapy (n = 978,420) were related to home remedies. Regarding body localization, most searches focused on the whole body (n = 109,050), followed by head (n = 89,360) and the genital area (n = 28,640).ConclusionsThe analysis of Google search data provides an overview of the populations’ interest regarding scabies. The analysis can detect local peaks and assess the relevance of scabies at individual localizations of the body. The study highlighted current possible shortcomings in the therapy of scabies. It also underlined the importance of improving awareness regarding scabies so that affected individuals can consult a doctor earlier for treatment.

Cyberspace-based urban networks: Visualising and exploring China's intercity interaction from a new perspective

The rapid urbanisation of China has increased public awareness of rising housing prices, particularly in large and medium-sized cities. The popularisation of the Internet in China has driven a shift in the public's attention to housing prices, moving from geographic space to cyberspace. Based on web search engine data using geotag, this paper attempts to create a better visualisation of the shift in the public's attention to housing prices from the perspectives of geographic space and cyberspace. Intercity interaction is also analysed in the context of cyberspace-based urban networks, through hybrid featured graphics. Specifically, data regarding public attention to housing prices are gathered (via queries data through Baidu Index) in 363 Chinese cities. The data are divided into a local city housing price attention index and an offsite city housing price attention index. Second, public attention flows are presented using origin-destination flows maps, chord diagrams and violin diagrams. These hybrid feature graphics show the interrelationship between cities and the complexity of public attention flows. This information will help us to reveal intercity interaction and the cyberspace-based urban network structure in China.

Can Google Trends data provide information on consumer’s perception regarding hotel brands?

Previous studies show that search engine query data is a valuable predictor for tourism demand forecasting. The goals of this study are to identify the current positions of hotels in the perception of the customer and to propose a method for practitioners to increase the visibility of consumer's mind perception of hotel brands. The study used volume of travel queries 30 hotel chains in the Turkey constructed from Google Trends and analyzed search query time series data (2014–2018). To visualize the position of brands was conducted social network analysis techniques. The results show that search engine query data regarding hotels reveal the positioning consumer’s mind of hotels. The study offers that Google Trends data is useful. In addition, the study proposes a method for practitioners. Tourism businesses could use search engine data to reveal its place in the consumer’s mind and change the consumer perception over the years.

China's Crude oil futures forecasting with search engine data

In this paper, we used the Baidu feed index to quantify the valuable information from the search engine, and proposed a China's crude oil futures price forecasting model with feed index. The empirical analysis confirms that the feed index, which characterizes the investor attention of participants in China's crude oil futures market, provides valuable information for internet concerns and investor sentiment, and has significant impact on China's crude oil futures price forecasting. We found that the time lag between the search query results and the crude oil futures price changes, which mostly lies within the shorter time range. The forecasting model with the search engine data would produce forecasts with superior performance.

Forecast Daily Tourist Volumes During the Epidemic Period Using Covid-19 Data, Search Engine Data and Weather Data

Forecast Daily Tourist Volumes During the Epidemic Period Using Covid-19 Data, Search Engine Data and Weather Data

A Novel Big Data Storage Reduction Model for Drill Down Search

Multi-level searching is called Drill down search. Right now, no drill down search feature is available in the existing search engines like Google, Yahoo, Bing and Baidu. Drill down search is very much useful for the end user to find the exact search results among the huge paginated search results. Higher level of drill down search with category based search feature leads to get the most accurate search results but it increases the number and size of the file system. The purpose of this manuscript is to implement a big data storage reduction binary file system model for category based drill down search engine that offers fast multi-level filtering capability. The basic methodology of the proposed model stores the search engine data in the binary file system model. To verify the effectiveness of the proposed file system model, 5 million unique keyword data are stored into a binary file, thereby analysing the proposed file system with efficiency. Some experimental results are also provided based on real data that show our storage model speed and superiority. Experiments demonstrated that our file system expansion ratio is constant and it reduces the disk storage space up to 30% with conventional database/file system and it also increases the search performance for any levels of search. To discuss deeply, the paper starts with the short introduction of drill down search followed by the discussion of important technologies used to implement big data storage reduction system in detail.

Use of Google Trends to Predict the Real Estate Market: Evidence from the United Kingdom

This paper demonstrates how Google Trends data can be used to improve real estate market predictions. Online searches produce valuable data that precede economic decisions. This study explores the usefulness of Google search engine data in predicting the real estate markets. The results indicate that Google data can be an additional source of information for investors and policymakers. This analysis adds to the existing literature that explores the role of behavioral factors in the decision-making process. Google Trends data are identified as an important predictor of real estate market prices and sales volume.

Now-casting Romanian migration into the United Kingdom by using Google Search engine data

Short-term forecasts of international migration are often based on data that are incomplete, biased, and reported with delays. There is also a scarcity of migration forecasts based on combined traditional and new forms of data.

AOL4PS: A Large-scale Data Set for Personalized Search

Personalized search is a promising way to improve the quality of Websearch, and it has attracted much attention from both academic and industrial communities. Much of the current related research is based on commercial search engine data, which can not be released publicly for such reasons as privacy protection and information security. This leads to a serious lack of accessible public data sets in this field. The few publicly available data sets have not become widely used in academia because of the complexity of the processing process required to study personalized search methods. The lack of data sets together with the difficulties of data processing has brought obstacles to fair comparison and evaluation of personalized search models. In this paper, we constructed a large-scale data set AOL4PS to evaluate personalized search methods, collected and processed from AOL query logs. We present the complete and detailed data processing and construction process. Specifically, to address the challenges of processing time and storage space demands brought by massive data volumes, we optimized the process of data set construction and proposed an improved BM25 algorithm. Experiments are performed on AOL4PS with some classic and state-of-the-art personalized search methods, and the experiment results demonstrate that AOL4PS can measure the effect of personalized search models.

Exam cheating and academic integrity breaches during the COVID-19 pandemic: An analysis of internet search activity in Spain

Academic institutions worldwide have had to adapt their methods of teaching and assessment due to COVID-19. In many Spanish universities exams moved online, affording students with opportunities to breach academic integrity that were not previously available to them. This paper uses the emerging research method of search engine data analysis to investigate the extent of requests for exam cheating information in Spain in the time period surrounding adjustments for the pandemic. The Internet data analytics technique is one that the paper proposes should be used more widely as an academic integrity research method.For this study, search engine activity data on exam cheating in Spain was collected and analysed for the five-year period between 2016 and 2020 inclusive. The data suggests that students are searching for information about ways to cheat in exams, including how to create cheat sheets. Most strikingly, the results show a significant increase in requests for information on cheating on online exams during the COVID-19 timeframe and the Spanish lockdown period. Based on the findings, academic institutions in other regions should be wary about the opportunities that their students have to commit exam fraud.

Forecasting sales using online review and search engine data: A method based on PCA–DSFOA–BPNN

Traditional sales forecasting methods are mainly based on historical sales data, which result in a certain lag. The relationship between sales volume and its influencing factors is intricate and often non-linear. In view of this, we propose a novel product forecasting method using online reviews and search engine data. Firstly, a dictionary-based sentiment analysis method is developed to convert the textual review concerning each attribute of the product into the corresponding sentiment score. And by combining the prospect theory and relevant online review data, sentiment indices in each period are calculated. Subsequently, data of product-related Baidu search words with different lag orders are collected and screened by time difference correlation analysis. Finally, the forecast model, PCA–DSFOA–BPNN, is constructed by combining the principal component analysis (PCA), the back propagation neural network (BPNN), and the improved fruit fly optimization algorithm (DSFOA), in which sentiment indices, Baidu search data, and historical sales volume are input data. Taking the monthly sales forecast of 14 automobile models as a case study, we observe that the proposed forecast method can effectively improve the forecast accuracy with good robustness.

Tourism demand nowcasting using a LASSO-MIDAS model

PurposeThis study aims to examine whether and when real-time updated online search engine data such as the daily Baidu Index can be useful for improving the accuracy of tourism demand nowcasting once monthly official statistical data, including historical visitor arrival data and macroeconomic variables, become available.Design/methodology/approachThis study is the first attempt to use the LASSO-MIDAS model proposed by Marsilli (2014) to field of the tourism demand forecasting to deal with the inconsistency in the frequency of data and the curse problem caused by the high dimensionality of search engine data.FindingsThe empirical results in the context of visitor arrivals in Hong Kong show that the application of a combination of daily Baidu Index data and monthly official statistical data produces more accurate nowcasting results when MIDAS-type models are used. The effectiveness of the LASSO-MIDAS model for tourism demand nowcasting indicates that such penalty-based MIDAS model is a useful option when using high-dimensional mixed-frequency data.Originality/valueThis study represents the first attempt to progressively compare whether there are any differences between using daily search engine data, monthly official statistical data and a combination of the aforementioned two types of data with different frequencies to nowcast tourism demand. This study also contributes to the tourism forecasting literature by presenting the first attempt to evaluate the applicability and effectiveness of the LASSO-MIDAS model in tourism demand nowcasting.