The Logical Dimensions, Technical Foundations, and Realization Forms of Big Data Analysis in Ideological and Political Education

Marx pointed out: “A science is only truly developed when it has reached the point where it can successfully employ mathematics” [1]1 This quotation is attributed to Marx. The source note “[1] (P8)” refers to a Chinese edition of Marx’s works, likely *Collected Works of Marx and Engels* (人民出版社), but the specific volume is not given in the original text. (P8). Big data technology enhances the ability of the discipline of ideological and political education to use mathematical methods to analyze and solve problems. In particular, the updating of equipment and algorithms enables states and characteristic changes such as human behavior, physiology, psychology, and emotions to be precisely captured, traced, described, calculated, and predicted. This greatly improves people’s abilities in qualitative analysis, quantitative analysis, computational simulation analysis, and the like. Promoting the application of big data analysis in ideological and political education, which takes people as its object, requires clarifying its logical dimensions, mastering its technical foundations, and exploring its realization forms.

I. The Logical Dimensions of Big Data Analysis in Ideological and Political Education

To promote the innovative development of ideological and political education in the era of big data, it is necessary to fully master and apply big data technology. Conducting big data analysis of people’s thoughts and behaviors requires not only having big data available to use, reflecting the effectiveness of data analysis, but also strengthening the logical construction of big data analysis within the discipline of ideological and political education.

1. From social data to people’s data. The traces generated by users’ operations and instructions on terminals constitute data. Big data possesses characteristics such as diversity, high velocity, value, and veracity. To move from massive data to data with value for ideological and political education, the focus of big data analysis needs to shift from holistic social data to specific people’s data. Each person’s data interacts with others’ data on the internet, forming multiple virtual social relationship networks. As the smallest data node within social data, people’s data can both reflect the different characteristics of individual thoughts and behaviors through its uniqueness and, by leveraging the co-construction and sharing nature of social data, reflect the common characteristics of group thoughts and behaviors. The complexity and diversity of social data are actually a virtual reflection of real society, and at root, they still manifest the complexity and diversity of individual thoughts and behaviors. In the process of big data analysis in ideological and political education, shifting the perspective from social data to people’s data requires being adept at extracting universal laws of social development from complex data, discovering the differentiated characteristics of individual needs from myriad data relationships, and matching social resources with the special needs of individuals. Grasping the basic logic of the shift from social data to people’s data means following the essential requirement of being people-centered, promoting a transformation of big data analysis thinking methods from the surface to the core, from the shallow to the deep, and from horizontal to vertical in-depth mining.

2. From known data to unknown data. Known data refers to limited-capacity small-sample data, which possesses basic characteristics such as finiteness, causality, and precision; unknown data refers to “big data,” which possesses characteristics such as infinity, correlation, and fuzziness. The complexity, diversity, and variability in the presentation forms of big data are driving people’s analytical logical thinking to shift from known data to unknown data. This requires people to move from known samples to unknown samples, specifically including shifting from precise samples to fuzzy samples, from causal samples to correlational samples, and from known samples to unknown samples. To make precise judgments about the “unknown,” it is necessary to move from “limited” small samples to “unlimited” large samples, from simple data statistics to the derivation of related data relationships, and from known facts to the exploration of unknown laws. By measuring various uncertain variables, big data analysis can provide decision-making and prediction services for people to solve various uncertainty problems, and this can also enhance the cognitive ability of ideological and political education regarding the variability of people’s thoughts and behaviors. On the one hand, based on acquiring and analyzing data on people’s thoughts and behaviors, especially conducting variable analysis of various influencing factors, regularities can be discovered in the overall characteristics of groups and the individualized characteristics of individuals, thereby continuously improving the initiative and targeting of ideological and political education. On the other hand, by mining and analyzing unknown data on people’s thoughts and behaviors on a larger scale and at a deeper level, especially conducting correlation analysis on elements that are difficult to grasp through objective observation and empirical judgment, trend predictions can be made for both environmental social influencing factors and individual’s own influencing factors, thereby continuously enhancing the discovery and judgment capabilities of ideological and political education.

3. From logical data to computational data. Logical data is relative to physical data. Physical data is all the data actually stored, while logical data is theoretical data obtained through logical analysis of physical data. Due to limitations in theoretical thinking and cognitive ability, it is difficult to exhaust the logical analysis of physical data; therefore, logical data cannot completely replace physical data. Computational data is data formed by algorithmically calculating physical data, aided by the continuously increasing computational power of big data algorithms. This not only enriches logical data in quantity and improves its quality but also makes the presentation of relationships in physical data more comprehensive, orderly, and regular. The transition from logical data to computational data implies the reproduction of theory, where calculated data results can be verified and repeated; it implies a shift in data from causal relationships to correlational relationships, where computation can discover correlational relationships hidden beneath the surface; it implies that the sorting of theoretical knowledge will no longer rely solely on empirical induction and summarization but can be obtained through verification and deduction. Promoting the in-depth exploration from logical data to computational data is a basic logic and important feature of big data analysis. Big data analysis in ideological and political education takes people’s physical data as the overall sample. Through algorithmic calculation, it can deeply discover data relationships between variables or phenomena of people’s thoughts and behaviors. Then, based on this, it conducts analysis and prediction that proceeds from one point to another and from the surface to the core, reflecting both the causal relationship analysis of logical data and the correlational relationship analysis of computational data. This continuously deepens the understanding of the universal correlational relationships and specific causal relationships existing in data on people’s thoughts and behaviors.

4. From quantitative data to qualitative data. Qualitative analysis and quantitative analysis are two basic methods for studying the development of things or social phenomena. Qualitative analysis studies the “quality” of things, mainly using methods such as logical deduction and empirical judgment to reveal the attribute characteristics and development laws of things; quantitative analysis studies the “quantity” characteristics of things, mainly using methods such as data statistics and mathematical calculation to reveal the characteristics, relationships, and changes of things. In specific scenarios or tasks, qualitative analysis and quantitative analysis can be used independently or comprehensively in a primary-auxiliary manner. Because people’s thoughts and behaviors are extremely complex and difficult to fully transform into quantifiable data, for ideological and political educators, it is still necessary to adhere to an analytical thinking method that primarily uses qualitative analysis, supplemented by quantitative analysis. Big data analysis in ideological and political education, on the one hand, cannot be separated from the thinking foundation of qualitative analysis and quantitative analysis—for example, qualitative analysis provides logical thinking for big data analysis, and quantitative analysis provides basic methods for big data analysis. On the other hand, using the thinking methods of qualitative analysis and quantitative analysis can form qualitative data and quantitative data. In particular, quantitative data based on behavioral characteristics can provide more objective, sufficient, and precise data support for qualitative data analysis based on ideological characteristics. Big data analysis in ideological and political education should enhance the breadth and depth of quantitative data analysis of people’s thoughts and behaviors, continuously strengthen the comprehensiveness, dynamism, and predictability of qualitative data analysis of people’s thoughts and behaviors, and thereby continuously promote data integration and method innovation among qualitative analysis, quantitative analysis, and big data analysis.

II. The Technical Foundations of Big Data Analysis in Ideological and Political Education

The logic of big data analysis is the prerequisite; technology is the foundation. Currently, research and application of big data analysis in areas such as network behavior discovery, physiological state capture, psychological state analysis, and emotional feature calculation are continuously deepening, providing a technical foundation for building digital analysis platforms for ideological and political education and mining the characteristics of people’s thoughts and behaviors.

1. Behavioral big data analysis based on network usage trajectories. As the scenario of “no one is off the internet, no time is off the internet, and no place is off the internet” [2]2 “No one is off the internet, no time is off the internet, and no place is off the internet” (无人不网、无时不网、无处不网): A Chinese political slogan emphasizing the ubiquity of internet access in contemporary society, often used in official discourse on digital governance and education. increasingly takes shape in human society, big data analysis of people’s online behavior has become an important method for ideological and political education to grasp the thoughts and behaviors of subjects and objects. Network behavioral big data analysis refers to the collection, organization, and analysis of data on users’ internet usage through the internet. Its main function is to form a visual presentation, data-based analysis, and value-oriented utilization of elements such as users’ online content, platforms, methods, and the corresponding internet usage needs, operational habits, and so forth. Magdalena Cantabella et al. used the Hadoop MapReduce framework to implement the Apriori algorithm, mining students’ operational events on the LMS platform, such as login behavior, use of different modules in the LMS, operations, and postings, and combining these with students’ basic information and grades to understand the potential correlation between students’ learning behavior and learning efficiency. As network behavioral big data analysis is increasingly widely applied to networked management services and digitalized production and daily life, the data source channels for ideological and political education will become increasingly abundant, and algorithmic analysis capabilities will also continue to strengthen, laying a good foundation for big data analysis in ideological and political education.

2. Physiological big data analysis based on the use of wearable devices. The application of human physiological electrical activity can be traced back to the medical field, later developing into multiple disciplines such as neurology, biology, and psychology, becoming an important pathway for scientists to observe human psychological activity and brain activity. Physiological condition big data analysis refers to a method that, with the support of Internet of Things technology, collects and analyzes data on skin electrical activity, electrocardiograms, brainwaves, skin conductance response, and respiratory frequency through users wearing biosensing equipment. Hussein Al Osman et al. used biofeedback serious games to collect subjects’ physiological electrical data and, using this as sample data, applied it to stress management applications. To achieve this goal, the researchers proposed a U-biofeedback reference model capable of testing and analyzing human mental stress states through continuous monitoring of heart rate variability, respiratory frequency, and activity levels. At the same time, by evaluating game feedback, they helped subjects adjust their own behavior to reduce stress levels. This research can provide a reference for ideological and political education to mine human emotional changes through physiological condition big data analysis. For example, in smart education and teaching spaces, with the help of lightweight physiological smart devices, educators can capture and analyze individuals’ physiological data such as eye movements, heart rate, and skin electrical activity to understand their physiological responses to different educational content, and thereby analyze their preferences for and acceptance of educational content, further providing decision-making support for optimizing teaching content and adjusting teaching methods.

3. Psychological big data analysis based on facial expression capture. A stable psychological state is the result of the combined action of multiple factors, and psychological measurement results based on scales can only represent psychological states within a certain time range rather than real-time dynamic measurements. Psychological state big data analysis refers to indirectly capturing data related to emotional states through direct measurement of facial expressions, personal postures, and the like, thereby providing data support for understanding the characteristics of psychological elements such as human perception, temperament, personality, interests, and motivation. Kiavash Bahreini et al., aiming to study human emotional states, used a facial expression recognition framework based on fuzzy logic algorithms to recognize and process human facial expressions, thereby reflecting human psychological activity. This framework uses the FURIA algorithm (Fuzzy Unordered Rule Induction Algorithm) to replace methods such as machine neural networks, Bayesian networks, and decision trees that require extensive computational processes, thereby calculating subjects’ psychological characteristics. Under the trend of promoting precision and personalization in ideological and political education, traditional measurement methods find it difficult to accurately grasp the individualized needs and preferences of educatees, requiring a comprehensive analysis linking psychology with external manifestations such as emotions and behaviors to obtain more accurate measurement results. For example, with the gradual popularization of technologies such as smart classrooms and online teaching, through the acquisition and analysis of videos and images in the online teaching environment based on a perspective of multi-source data fusion, students’ facial expressions and body postures can be accurately captured, enhancing the cognitive ability regarding students’ psychological characteristics in the process of networked learning and daily life.

4. Emotional big data analysis based on audio and video information processing. Under the trend where “human-machine” interaction is increasingly replacing “human-human” interaction, the issue of human emotional deficiency in networked production and daily life has also become a key area of interdisciplinary research in the humanities. Affective computing big data analysis is a method that conducts comprehensive mining and analysis of various types of data to study the emotional characteristics and expression patterns of people in online social activities. With the impetus of modern information technology, affective computing big data analysis for humans can mine relevant information from different data sources such as video data, text data, audio data, and physiological data, and, based on certain natural language algorithms, achieve the calculation of emotional dimensions. For example, Taguchi et al. used speech recognition technology to explore the treatment of patients with depression, using the OpenSMILE component to investigate the relationship between Mel-frequency cepstral coefficients and biomarkers of depression patients, as well as collecting and extracting features from patients’ speech during the treatment process, and comparing the voice characteristics of patients with those of normal individuals. Introducing affective computing big data analysis into the field of online education and teaching in ideological and political education can provide effective decision-making support for solving the problem of lack of emotional communication in online teaching models.

III. Implementation Forms of Big Data Analysis in Ideological and Political Education

In ideological and political education, it is necessary to fully leverage the advantages of big data technology, such as holistic perspectives, dynamic patterns, correlational logic, and mining techniques, and actively construct a four-dimensional big data analysis platform based on environmental variables, knowledge cognition, emotional identification, and behavioral tendencies.

1. Big data analysis based on environmental variables. The environment itself is a complex system, whose constituent elements can be roughly divided into family, school, community, and network environments. Ideological and political education takes environmental variables as the object of analysis, using big data technology to deeply mine and utilize environment data centered on people, achieving a transformation from macro-level analysis of the social environment to micro-level analysis of the environment affecting individual growth. The subtle “immersion-type” education of the family accompanies a person throughout their life. Through registered data records, it is possible to understand the family environment of educatees, such as the occupations, educational levels, political statuses, and family incomes of main family members. The school environment includes teaching infrastructure, the surrounding teaching environment, educational teaching resources, as well as teacher-student relationships, organizational relationships, and peer relationships. Data sources for the school environment can be existing measurement data, such as data related to physical or natural factors like parameters of infrastructure, teaching duration, classroom temperature and humidity; they can also be network data from electronic educational resource libraries; or data in other forms, such as data from qualitative analysis of interview materials, organizational documents, and written evaluations. The community environment undertakes functions such as the allocation of educational resources and the shaping of values and culture; its data sources are mainly based on local chronicles, local archives, local media, as well as local economic development, cultural activities, and traditional customs. The network environment is the material and cultural foundation for human interaction across time and space through the internet. By collecting data on the network environment, it is possible to understand educatees’ participation, activity levels, and expressiveness in cyberspace, thereby revealing the degree of influence of network environmental variables on them.

By collecting, organizing, and comprehensively computing data on the above environmental variables, the optimization and construction of a “dynamic environment” for ideological and political education can be promoted. First, data analysis is conducted on the infrastructure construction and service guarantee levels on which ideological and political education depends, so as to promote the dynamic adjustment of the hardware environment for ideological and political education. For example, by conducting an index evaluation of a school’s building area, functional layout, ecological environment, and the degree to which various infrastructure facilities meet student needs, an analytical index reflecting the strength of the hardware environment’s supporting role in ideological and political education can be formed, and the hardware environment can be adjusted based on this index. Second, big data analysis is conducted on the different environments and their variable factors during the growth process of the educated subjects, so as to promote the dynamic adjustment of the growth environment for the objects of ideological and political education. For example, by evaluating their preferred ways of living, activities, and socializing, an analytical index reflecting the role of environmental variable factors in ideological and political education at different stages and in different fields can be formed, and based on this index, a good campus culture can be built, campus cultural and sports activities can be organized, and so on. Third, big data analysis is conducted on the literacy and ability, working methods, and management mechanisms of the educational subjects, so as to promote the dynamic adjustment of the working environment for the ideological and political education workforce. For example, by evaluating the age, gender, work style, and interaction with the educated subjects of the ideological and political education workforce, an analytical index reflecting the work level of the ideological and political education workforce can be formed, and based on this index, the composition, style, and ratio of the workforce can be adjusted to maximize educational effectiveness.

2. Big Data Analysis Based on Knowledge Cognition. Cognition is an important component of human thinking; the level of cognition determines an individual’s ability to accept knowledge in ideological and political education and also affects an individual’s willingness to accept such knowledge. Taking knowledge cognition as the object of analysis, ideological and political education uses big data technology to conduct in-depth analysis of the data from the process of human knowledge learning, realizing a transformation from the analysis of knowledge learning behavior to the analysis of knowledge cognitive ability. Among these, knowledge acceptance ability is a manifestation of an individual’s ability to recognize and remember; its data mainly come from online and offline sources, such as traditional test and assessment results, and data from cloud classrooms and information management systems. Knowledge acceptance willingness is an individual’s initiative and enthusiasm for learning and accepting objective knowledge; with the help of big data technology, it is possible to capture an individual’s learning habits, learning methods, etc., to understand their learning willingness. Knowledge application ability is an individual’s ability to re-cognize and innovatively apply knowledge; by capturing speech records in online communities and using keywords to screen and analyze relevant content, big data technology can judge the knowledge application ability of different individuals.

In teaching processes across different scenarios and levels, big data technology can help educators achieve personalized, intelligent, and precise educational evaluation, so as to construct a “dynamic teaching scenario.” First, personalization of teaching content is achieved. That is, educators form analytical indices reflecting different educational contents through big data analysis that can reflect the content preferred by the educated subjects, thereby achieving personalization of teaching content. For example, by sorting and analyzing students’ search records on the internet, their points of interest and needs in learning ideological and political theoretical knowledge can be accurately grasped, thereby dynamically adjusting the content of ideological and political education. Second, precision of teaching pathways is achieved. That is, through comparative analysis of data from the application of different educational methods, analytical indices reflecting different educational methods are formed. For example, for education on the same theme, by comparing the analytical indices formed by different educational methods such as classroom lectures, watching documentaries, and practical visits, educators can accurately grasp the degree of fit between different educational methods and the receptive methods of the educated subjects, thereby dynamically adjusting teaching methods. Third, intelligence of teaching procedures is achieved. That is, through comparative analysis of data from the application of different educational procedures, analytical indices reflecting different educational procedures are formed. For example, for the same work task, by conducting differential analysis of the big data formed by different spatiotemporal arrangements and procedural arrangements, educators can accurately grasp the degree of alignment between different implementation processes and the cognitive laws of the educated subjects, thereby dynamically adjusting the arrangement of teaching procedures.

3. Big Data Analysis Based on Emotional Identification. The support of emotion is the bridge for the transformation from “knowing” to “doing.” [2]2 “No one is off the internet, no time is off the internet, and no place is off the internet” (无人不网、无时不网、无处不网): A Chinese political slogan emphasizing the ubiquity of internet access in contemporary society, often used in official discourse on digital governance and education. According to relevant theories of emotion research, ideological and political education can conduct big data analysis from dimensions such as emotional expression, emotional tendency, and emotional response. Taking emotional identification as the object of analysis, ideological and political education can use big data technology to conduct comprehensive analysis of data on human physiology, behavior, etc., realizing a transformation from empirical judgment of human emotional changes to intelligent analysis. Among these, emotional expression refers to the transmission of emotions by the educated subjects through speech, expressions, sounds, etc. By analyzing data such as text, audio, and video from the educated subjects in classrooms, social media, or other platforms, educators can infer individual emotional expression. Emotional tendency refers to the emotional orientation of the educated subjects toward a specific object, theme, or event. Educators can scrape data from the remarks of the educated subjects on social media, learning platforms, or other online communication platforms, and use certain algorithms to analyze the variable factors of individual emotions and their different effects. Emotional response refers to the physiological and psychological reactions of the educated subjects when facing stimuli. Educators can use big data technology to analyze physiological signal data of the educated subjects in laboratories or virtual reality environments, such as electroencephalography (EEG), heart rate variability (HRV), skin conductance response (SCR), etc., to evaluate individual emotional responses to external stimuli.

By collecting, organizing, and comprehensively computing data on the above emotional dimensions, ideological and political education can construct a “smart psychological counseling scenario.” This scenario can help educators achieve element awareness, dynamic mastery, and identification assessment. First, element awareness is achieved. Educators conduct big data analysis on the personality, preferences, and psychological quality of the educated subjects, forming analytical indices that reflect the characteristics of their emotional changes, thereby integrating resources and optimizing educational methods to fully mobilize the positive emotions of the educated subjects. Second, dynamic mastery is achieved. Educators analyze data on the emotional changes of the educated subjects in different scenarios, forming analytical indices that can identify and predict changes in their psychological expectations, and scientifically predict possible psychological disorders or crisis events. Third, identification assessment is achieved. Educators analyze data on the emotional changes and psychological activities of the educated subjects toward the same thing in different scenarios, forming an analytical index of their relatively stable emotional identification with this thing, understanding their emotional identification status with this thing, and thereby providing an objective basis for analyzing and deepening their theoretical identification.

4. Big Data Analysis Based on Behavioral Tendencies. Scientific and reasonable data classification analysis can enhance the pertinence and efficiency of grasping behavioral characteristics. The collection and analysis of big data in ideological and political education can be carried out according to three dimensions: behavioral habits, behavioral transformation, and degree of participation. Taking human behavioral tendencies as the object of analysis, ideological and political education uses big data technology to conduct comprehensive analysis of human physiological and behavioral data, thereby achieving a transformation from known behavioral patterns to predictions of unknown behavioral trends. Among these, behavioral habits are relatively fixed behavioral patterns formed by individuals after repeated practice over a period of time. By analyzing various types of network data of the educated, including front-end usage data of network platforms, back-end interaction data of network platforms, and physiological data generated by sensor devices, educators can understand their behavioral trajectories and data characteristics. Behavioral transformation is the degree of consciousness with which the educated transform the theoretical knowledge they have learned into practical action. Through behavioral records on network platforms and social media data, educators can understand the behavioral tendencies of the educated in political participation, social participation, cultural participation, etc., thereby enhancing their ability to judge these behavioral tendencies. Degree of participation refers to the breadth and depth of an individual's participation in activities or events. Through big data analysis of the frequency of the educated's participation in activities and the quality and efficiency of task completion, educators can understand the breadth and depth of their participation in activities.

Big data technology can capture the behavior of the educated in different scenarios, thereby constructing more comprehensive, more accurate, and deeper "student profiles." [4]4 "Student profiles" (学生画像) is a term in educational data mining referring to a comprehensive data-driven representation of a student's attributes, behaviors, and potential development trajectories. First, achieving a more comprehensive behavioral profile. That is, according to different analytical needs of ideological and political education, cluster analysis is carried out according to the corresponding logic, forming analytical indices that reflect the behavioral characteristics of the educated, providing a convenient channel for timely and accurately grasping their ideological and behavioral characteristics and changes. Second, achieving a deeper ideological profile. That is, through logical construction and data analysis of the corresponding relationship between data patterns and ideological characteristics, analytical indices that present ideological characteristics through data relationships are formed, which is conducive to analyzing and speculating on differential changes in their ideology, thereby enhancing the value of behavioral data analysis. Third, achieving a more accurate future profile. That is, through data analysis of data associated with the characteristics of behavioral changes of the educated, trend analysis indices reflecting tendencies toward a specific problem or behavior are formed, which can greatly improve the ability to discover and judge students' future behavioral tendencies.

(Author Biographies: Liu Hongda, Professor and Doctoral Supervisor at the School of Marxism, Central China Normal University; Liao Yinwei, Doctoral Student at the School of Marxism, Central China Normal University)

4. Big data analysis based on behavioral tendencies. Scientific and reasonable data classification analysis can enhance the pertinence and efficiency of grasping behavioral characteristics. The big data collection and analysis of ideological and political education can be carried out according to three dimensions: behavioral habits, behavioral transformation, and degree of participation. Taking human behavioral tendencies as the object of analysis, ideological and political education uses big data technology to conduct comprehensive analysis of human physiological and behavioral data, thereby realizing a shift from known behavioral patterns to the prediction of unknown behavioral trends. Among these, behavioral habits are relatively fixed behavioral patterns formed by individuals after repeated practice over a period of time. By analyzing various types of network data of the educated, including front-end usage data, back-end interaction data, and physiological data generated by sensor devices, educators can understand their behavioral trajectories and data characteristics. Behavioral transformation is the degree of consciousness with which the educated transform the theoretical knowledge they have learned into practical action. Through behavioral records on network platforms and social media data, educators can understand the behavioral tendencies of the educated in political participation, social participation, cultural participation, etc., thereby enhancing their ability to judge these behavioral tendencies. Degree of participation refers to the breadth and depth of an individual’s participation in activities or events. By conducting big data analysis on the frequency of the educated’s participation in activities and the quality and efficiency of task completion, educators can understand the breadth and depth of their participation.