Search

This study examined whether the behaviour of Internet search users obtained from Google Trends contributes to the forecasting of two Australian macroeconomic indicators: monthly unemployment rate and monthly number of short-term visitors. We assessed the performance of traditional time series linear regression (SARIMA) against a widely used machine learning technique (support vector regression) and a deep learning technique (convolutional neural network) in forecasting both indicators across different data settings. Our study focused on the out-of-sample forecasting performance of the SARIMA, SVR, and CNN models and forecasting the two Australian indicators. We adopted a multi-step approach to compare the performance of the models built over different forecasting horizons and assessed the impact of incorporating Google Trends data in the modelling process. Our approach supports a data-driven framework, which reduces the number of features prior to selecting the best-performing model. The experiments showed that incorporating Internet search data in the forecasting models improved the forecasting accuracy and that the results were dependent on the forecasting horizon, as well as the technique. To the best of our knowledge, this study is the first to assess the usefulness of Google search data in the context of these two economic variables. An extensive comparison of the performance of traditional and machine learning techniques on different data settings was conducted to enable the selection of an efficient model, including the forecasting technique, horizon, and modelling features. [ABSTRACT FROM AUTHOR]

Heat stroke is a serious heat‐related health outcome that can eventually lead to death. Due to the poor accessibility of heat stroke data, the large‐scale relationship between heat stroke and meteorological factors is still unclear. This work aims to clarify the potential relationship between meteorological variables and heat stroke, and quantify the meteorological threshold that affected the severity of heat stroke. We collected daily heat stroke search index (HSSI) and meteorological data for the period 2013–2020 in 333 Chinese cities to analyze the relationship between meteorological variables and HSSI using correlation analysis and Random forest (RF) model. Temperature and relative humidity (RH) accounted for 62% and 9% of the changes of HSSI, respectively. In China, cases of heat stroke may start to occur when temperature exceeds 36°C and RH exceeds 58%. This threshold was 34.5°C and 79% in the north of China, and 36°C and 48% in the south of China. Compared to RH, the threshold of temperature showed a more evident difference affected by altitude and distance from the ocean, which was 35.5°C in inland cities and 36.5°C in coastal cities; 35.5°C in high‐altitude cities and 36°C in low‐altitude cities. Our findings provide a possible way to analyze the interaction effect of meteorological variables on heat‐related illnesses, and emphasizes the effects of geographical environment. The meteorological threshold quantified in this research can also support policymaker to establish a better meteorological warning system for public health. Plain Language Summary: The impact of extreme heat events on population has become of urgent public health concern. In China, the real mortality and morbidity data are not publicly available and data of some diseases are not enough recorded. Thus, it is difficult to build a solid relationship between heat strokes and meteorological variables in large spatial scale. Internet search data, highly correlated with real heat stoke cases, was adopted as a new data set in our research instead of the heat‐related morbidity data to investigate the relationship between heat strokes and multiple meteorological variables. According to Random forest model, temperature and relative humidity (RH) were identified as the most important factor, which mainly affected the severity of heat strokes. We quantified the meteorological threshold that affected the severity of heat strokes as well. Heat strokes may start to occur when temperature exceeds 36°C and RH exceeds 58% in China. The temperature of this threshold is higher in the south or low‐latitude region or coastal region of China. This work provides new insights for health research and help to timely alert the public in adverse weather conditions. Key Points: Internet search data could help to evaluate the health impact in large scale to address the data shortageIncreasing temperature and relative humidity (RH) led to sharply increased heat stroke, particularly when maximum temperature exceeds 36°C and RH exceeds 58%Meteorological threshold was affected by geographical factors like altitude, latitude and distance from ocean [ABSTRACT FROM AUTHOR]

Before making travel plans, people often use the Internet to collect relevant information to help themselves make better decisions. Among the numerous information search channels, Internet search engine is used by the vast number of travelers because of its low cost and high efficiency. To a large extent, Internet search behavior is the external manifestation of users' psychological activities, reflecting their concerns, needs and preferences. Therefore, Internet search data can reflect the air passenger demand information to a certain extent. In this manuscript, a novel decomposition ensemble model is proposed to discuss the role of Internet search data in air passenger demand forecasting. In the empirical study, the relevant data of Shanghai Pudong International Airport and Beijing Capital International Airport are taken as samples. The results show that the proposed forecasting model can integrate the advantages of decomposition-ensemble strategy and deep learning algorithm, and achieve more accurate and reliable prediction results than all benchmark models. This further indicates that adding Internet search data into the forecasting model can effectively improve the prediction performance of air passenger demand, and can provide scientific and reliable decision support for air transport management. [ABSTRACT FROM AUTHOR]

To avoid the effects of systemic financial risks caused by extreme fluctuations in housing price, the Chinese government has been exploring the most effective policies for regulating the housing market. Measuring the effect of real estate regulation policies has been a challenge for present studies. This study innovatively employs big data technology to obtain Internet search data (ISD) and construct market concern index (MCI) of policy, and hedonic price theory to construct hedonic price index (HPI) based on building area, age, ring number, and other hedonic variables. Then, the impact of market concerns for restrictive policy, monetary policy, fiscal policy, security policy, and administrative supervision policy on housing prices is evaluated. Moreover, compared with the common housing price index, the hedonic price index considers the heterogeneity of houses and could better reflect the changes in housing prices caused by market supply and demand. The results indicate that (1) a long-term interaction relationship exists between housing prices and market concerns for policy (MCP); (2) market concerns for restrictive policy and administrative supervision policy effectively restrain rising housing prices while those for monetary and fiscal policy have the opposite effect. The results could serve as a useful reference for governments aiming to stabilize their real estate markets. [ABSTRACT FROM AUTHOR]

With the increasing popularity of tourism activities, the forecasting of tourist volume has become an important research issue in the field of tourism management. However, the traditional statistical data cannot reflect the changes in tourism demand in real time. In order to make up for this shortcoming, scholars have found that web search data and big data technologies can provide a new way to forecast tourism demand which can expose user behavioral intentions in real time. Accordingly, this paper tries to make a prediction of the number of China inbound foreign tourists based on Google Trends data, and by applying Random Forest (RF) model to this task, a higher prediction accuracy has been achieved. [ABSTRACT FROM AUTHOR]

In recent years, visitation to U.S. National Parks has been increasing, with the majority of this increase occurring in a subset of parks. As a result, managers in these parks must respond quickly to increasing visitor-related challenges. Improved visitation forecasting would allow managers to more proactively plan for such increases. In this study, we leverage internet search data that is freely available through Google Trends to create a forecasting model. We compare this Google Trends model to a traditional autoregressive forecasting model. Overall, our Google Trends model accurately predicted 97% of the total visitation variation to all parks one year in advance from 2013 to 2017 and outperformed the autoregressive model by all metrics. While our Google Trends model performs better overall, this was not the case for each park unit individually; the accuracy of this model varied significantly from park to park. We hypothesized that park attributes related to trip planning would correlate with the accuracy of our Google Trends model, but none of the variables tested produced overly compelling results. Future research can continue exploring the utility of Google Trends to forecast visitor use in protected areas, or use methods demonstrated in this paper to explore alternative data sources to improve visitation forecasting in U.S. National Parks. • Search engine data is useful for predicting visitation to park units the next year. • Overall, using Google Trends performs better than an autoregressive model. • Using Google Trends to forecast visitation is useful for some, but not all, parks. • Utility of Google Trends is weakly correlated with higher surrounding populations. [ABSTRACT FROM AUTHOR]

We use Google searches of the word ‘mortgage’ to explain monthly housing transactions in the Netherlands in the period from 2004 until 2015. Our estimates indicate that Google searches of the previous months are significantly positively associated with housing transactions in the current month. This shows evidence that Internet search data can provide information about real market behaviour. [ABSTRACT FROM AUTHOR]

Previous studies have revealed that Internet search data is a new source of data that can be used to predict the stock market. In this new, data-driven research field, choosing a method for preprocessing data is crucial to achieving accurate prediction performance. This paper proposes a preprocessing method of Internet search data: composite leading search index (CLSI), which is composed of three steps: (a) keyword selection, (b) time difference measurement, and (c) leading index composition. We demonstrate the validity of CLSI by comparing this method's results with the results from search volume index (SVI), which is most commonly used in previous literatures. We build a time series model (TS) with error correction and support vector regression (SVR) for stock trend prediction, and combine into four models for comparison: SVI-TS, CLSI-TS, SVI-SVR, and CLSI-SVR. We test these four models in the context of the Chinese stock market, which interests more and more investors nowadays, and analyzed results in nine datasets: stable periods, peak periods and trough periods of Shanghai Composite Index, Shenzhen Composite Index, and Hushen 300 index respectively. The results show that using TS and SVR as forecasting models, CLSI performs better than SVI on majority of the test dataset while has almost the same performance with that of SVI on the remaining test dataset. It is to some extent convincing that CLSI is a more efficient preprocessing method of Internet search data for stock trend prediction. [ABSTRACT FROM AUTHOR]

The alt-right is a white supremacist social movement that operates primarily online. Its broader constituency has not been studied systematically. Participants in white supremacist movements tend to join in response to threats to their social and economic status. Quantitative work suggests they come primarily from working- and lower-middle class backgrounds. Alt-right leadership, however, argues their movement successfully mobilizes a more affluent population of college-educated professionals. In this paper, we examine predictors of county-level Internet search volume for alt-right content. Results indicate that counties with larger percentages of college graduates, of highly educated non-white and immigrant groups, and higher poverty levels for college graduates tend to have a higher search volume for alt-right content. We interpret this as evidence that the alt-right appeals to college-educated whites experiencing real or perceived threats to their economic and social status. [ABSTRACT FROM AUTHOR]