The Social Media Big Data Analysis for Demand Forecasting in the Context of Globalization: Development and Case Implementation of Innovative Frameworks

Introduction

With the continuous deepening of globalization, social media has become an integral part of people's daily life. The rapid increase in user activity and information exchange on social media platforms can provide companies with a larger market and more opportunities to obtain user data. In this increasingly competitive market environment, these data can help companies better understand market demand and consumer behavior, enhance their competitiveness and economic efficiency, and increase the market share (Meng et al., 2022; Rajković et al., 2021). However, these methods suffer from information lag and limitations, often making it difficult to make timely and accurate predictions of rapidly changing market demands. To better predict consumer demand and behavior, companies need to master and analyze a large amount of data (Chen et al., 2021). With the popularization of the Internet and social media, the application of big data is becoming increasingly widespread. Social media has become a highly potential source of data. Big data analysis techniques are used to analyze user behavior and speech in social media. Enterprises can better understand consumers' needs and preferences and adjust product and service strategies (Lei et al., 2022; Tajpour et al., 2023). Therefore, how to use social media big data to predict user needs accurately has become the focus of many scholars in related fields.

With the rise and popularity of social media, more and more people are starting to express their opinions and needs through social media. Using big data on social media for demand forecasting has become a new approach, and many scholars have conducted relevant research. Feng et al. (2021) studied environmental governance startups and pointed out that the information sharing of venture capital networks and social media positively correlated with investment performance. In social media, the degree of enterprise risk information sharing was increasing. Based on the native big data of the Internet of Things, Chen and Du (2022) elaborated on the blocking process of the evolution of Online Public Sentient (OPS) and utilized artificial intelligence (AI) technology and big data to simulate and predict the development of OPS. The results indicated that Adam's optimized Long Short-Term Memory (LSTM) neural network model could predict the heat of OPS in dynamic evolution with high prediction accuracy. Dai (2022) improved the existing CNN and applied it to social media for predicting user demand in the stock market. The results showed that the proposed improved convolutional neural network had significant advantages in prediction accuracy based on specific data from the Chinese stock market for empirical analysis. Liu and Chen (2023) proposed a new potential feature topic model and introduced a time series model to establish a topic evolution network. The high-frequency words from three periods in social media were compared and analyzed. The results showed that the topic model found the evolution law of AI domain knowledge structure. The research of these scholars shows that the digital transformation of enterprises is the strategic demand of enterprise development at present and that it plays an important role in promoting economic benefits. However, there are relatively few relevant studies on user demand prediction in enterprise products from the perspective of social media big data, such as on the specific influencing factors of the development direction of social media big data.