Research on the Strategy of Using Data Visualization Technology to Enhance the Effect of Ideological and Political Education of Students in Colleges and Universities

Big data has become the current cutting-edge research direction in the field of information technology, and it has been integrated into various fields of society, becoming one of the key driving forces to promote the development of social innovation. Data visualization technology aims to transform complex and massive data into intuitive graphs, charts or maps to reveal patterns, trends and associations in the data by visual means and help users better understand and analyze the data [1-4]. The technology originated in the field of scientific computing and statistical analysis, and in recent years, with the development of computer graphics, big data analysis technology, human-computer interaction interface and other fields, as well as the integration of artificial intelligence and machine learning algorithms, the data visualization technology continues to evolve, and has become an important part of data science. Data visualization technology has significant features such as visibility, multidimensionality and interactivity [5-8]. Visualization refers to the display of data through images, charts, maps and other visual forms, so that users can intuitively observe, understand and analyze the data. Multidimensionality refers to the fact that users can classify, sort, combine and display the data information with different conditions and dimensions as the standard, so as to realize the multi-angle observation and analysis of data information [9-12]. Interactivity means that users can perform operations such as scaling, filtering, drill-down, etc. to interact with the visualized data and realize in-depth analysis and exploration of data. In the teaching of ideological and political education for students in colleges and universities, the application of data technology is becoming more and more widespread, which effectively improves the teaching efficiency and level [13-16]. As one of the basic courses in colleges and universities, the introduction of data visualization technology in ideological and political courses helps to overcome the shortcomings of traditional teaching and improve teaching efficiency and level. Therefore, exploring the application strategy of data visualization technology in the teaching of ideological and political education is of great significance in promoting teaching innovation and improving teaching quality [17-20].

This study constructs a visual learning analytics tool with a design framework that integrates the theoretical foundations of visual learning analytics, information visualization, and educational data narratives. The strategy was embedded into an ideological and political online education and learning platform, and a multifaceted design of the teaching experiment program was carried out in terms of implementation of teaching, data collection, and visualization of the results in order to evaluate the impact of the technology on the effectiveness of education. In the use of CART regression tree, a classification study is conducted based on students’ behaviors in the classroom to further validate the role played by data visualization and ideological and political education online learning platform.

2

Framework for a visual learning analytics tool based on civic education

2.1

Visual Learning Analytics

Visual learning analytics is a method that uses analytics and visualization techniques to aid in the understanding of complex learning phenomena and solve complex learning problems.

Visual learning analytics can be used as a novel methodology to support the understanding of complex learning problems/phenomena by providing open-ended exploratory and validation capabilities that incorporate human experience, intelligence, creativity, and flexibility. This iterative development process is usually non-linear and orderly, which can be presented through a visual learning analytics process model. The visual learning analytics process model is shown in Figure 1. The process model highlights the ability of the user to guide the visual analytics process by interacting to optimize the input parameters of the analytics process or focusing on different parts of the data to validate the generated hypotheses or extracted insights.

2.2

Information visualization

Information visualization refers to interactive data exploration and visual representation methods based on visual perception, which can be used to help people better discover patterns, relationships, and trends in data. In this study, two mainstream theories were selected as theoretical references for design principles.

2.2.1

Cognitive load theory

Cognitive load theory [21] explains the cognitive overload that people face with limited working memory and how it affects their cognitive processes. Visual learning analytics is the process of processing and comprehending large amounts of information, and the theory can be utilized to guide visualization design strategies. Specifically, this includes removing and reducing redundant information, optimizing visualization layouts, reducing the amount of information users need to search for and move around the screen, and reducing external cognitive load by allowing users to focus on information that is relevant to their goals. Sorting out information hierarchies and displaying them in layers, so that users understand them layer by layer, reducing internal cognitive load, etc.

2.2.2

Dual coding

Dual coding theory [22] states that human beings use two different coding and processing systems when processing information, i.e., a linguistic-based coding and processing system and a non-linguistic-based coding and processing system. Dual coding theory has been applied to design visualization systems with good readability by using strategies that include: using multiple visual elements to provide more intentional classes of coding. Providing verbal coding through text labels, cue messages, and sounds. Using coding that combines visual and verbal elements, such as using data labels to highlight important information.

2.3

CART-based Visualization Generation Generation

CART regression tree generation [23] algorithm:

Input: training dataset D.

Output: regression tree f(x).

Construct a binary decision tree by recursively dividing each region into two subregions in the training data set and deciding the output value of each subregion: 1)

Select the best cut-off variable j and the best cut-off point s and solve: (1) $\min_{j, s} [\min_{c_{1}} \sum_{x_{i} \in R_{1} (j, s)} {(y_{i} - c_{1})}^{2} + \min_{c_{2}} \sum_{x_{i} \in R_{2} (j, s)} {(y_{i} - c_{2})}^{2}]$ $${\min _{j,s}}\left[ {{{\min }_{{c_1}}}\sum\limits_{{x_i} \in {R_1}(j,s)} {{{\left( {{y_i} - {c_1}} \right)}^2}} + {{\min }_{{c_2}}}\sum\limits_{{x_i} \in {R_2}(j,s)} {{{\left( {{y_i} - {c_2}} \right)}^2}} } \right]$$

Traverse j and for a fixed cut variable j scan the cut point s to find the pair (j, s) for which the above equation reaches its minimum value. 2)

For the selected pair (j, s) divide the region and decide the corresponding output value: (2) $R_{1} (j, s) = {x | x^{(j)} \leq s}, R_{2} (j, s) = {x | x^{(j)} > s}$ $${R_1}(j,s) = \left\{ {x|{x^{(j)}} \leq s} \right\},{R_2}(j,s) = \left\{ {x|{x^{(j)}} > s} \right\}$$ (3) ${\hat{c}}_{m} = \frac{1}{N_{m}} \sum_{x_{i} \in R_{m} (j, s)} y_{i}, x \in R_{m}, m = 1, 2$ $${\hat c_m} = \frac{1}{{{N_m}}}\sum\limits_{{x_i} \in {R_m}(j,s)} {{y_i}} ,x \in {R_m},m = 1,2$$

3)

Continue with steps 1) 2) for both regions until the condition is satisfied.

4)

Divide the input space into M cell R₁, R₂, …, R_M and generate a decision tree: (4) $f (x) = \sum_{m = 1}^{M} \hat{c_{m}} I (x \in R_{m})$ $$f(x) = \sum\limits_{m = 1}^M {\widehat {{c_m}}} I\left( {x \in {R_m}} \right)$$

Subsequently CART pruning is performed, the CART pruning algorithm is divided into two steps: 1)

Pruning to form a sequence of subtrees

During the pruning process, the loss function of the subtree is calculated: (5) $C_{α} (T) = C (T) + α | T |$ $${C_\alpha }(T) = C(T) + \alpha \left| T \right|$$

where T is an arbitrary subtree, C(T) is the prediction error (e.g., squared error minimization) of the training data, $| T |$ $$\left| T \right|$$ is the number of leaf nodes of the subtree, α ≥ 0 is a parameter, and C_α(T) is the overall loss of the subtree $| T |$ $$\left|T\right|$$ when the parameter is α. Parameter α weighs the complexity of the model and the goodness of fit of the training data.

For fixed α, there must exist a subtree that minimizes the loss function C_α(T), which we denote by T_α. Specifically, pruning starts from the overall tree T₀.

For any internal node t of T₀, the loss function for a single node tree with t is: (6) $C_{α} (t) = C (t) + α$ $${C_\alpha }(t) = C(t) + \alpha$$

The loss function of subtree T_t with t as the root node is: (7) $C_{α} (T_{t}) = C (T_{t}) + α | T_{t} |$ $${C_\alpha }\left( {{T_t}} \right) = C\left( {{T_t}} \right) + \alpha \left| {{T_t}} \right|$$

When α = 0 or α is sufficiently small, there is an inequality: (8) $C_{α} (T_{t}) < C_{α} (t)$ $${C_\alpha }\left( {{T_t}} \right) < {C_\alpha }(t)$$

When α increases, there will be: (9) $C_{α} (T_{t}) = C_{α} (t)$ $${C_\alpha }\left( {{T_t}} \right) = {C_\alpha }(t)$$

When α is any larger, inequality (8) is reversed. As soon as $α = \frac{C (t) - C (T_{t})}{| T_{t} | - 1}$ $$\alpha = \frac{{C(t) - C\left( {{T_t}} \right)}}{{\left| {{T_t}} \right| - 1}}$$, T_t and t will have the same value of the loss function, but at this point t has fewer nodes, so t is preferable to T_t, so prune T_t.

So for each node t in T₀, calculate: (10) $g (t) = \frac{C (t) - C (T_{t})}{| T_{t} | - 1}$ $$g(t) = \frac{{C(t) - C\left( {{T_t}} \right)}}{{\left| {{T_t}} \right| - 1}}$$

It indicates the extent to which the overall loss function is reduced after pruning. Cut g_t the smallest T_t in T₀ and take the resulting subtree as T₁, while setting the smallest g_t as α₁. T₁ is then the optimal subtree for interval $[α_{1}, α_{2})$ $$\left[ {{\alpha _1},{\alpha _2}} \right)$$. 2)

Select the optimal subtree

Specifically, the prediction error of each subtree in the subtree sequence T₀, T₁, …, T_n is tested using an independent test set. The resulting decision tree with the smallest prediction error is considered to be the optimal decision tree. In the subtree sequence, each subtree T₁, T₂, …, T_n corresponds to a parameter α₁, α₂, …, α_n. Therefore, when the optimal subtree T_k is determined, the corresponding α_k is also determined, i.e., the optimal decision tree T_α is obtained.

CART Regression tree pruning algorithm:

Input: CART Decision tree generated by regression tree T₀.

Output: optimal decision time T_α. 1)

Set k = 0, T = T₀.

2)

Set α = + ∞.

3)

Calculate $C (T_{t})$ $$C\left( {{T_t}} \right)$$, $| T_{t} |$ $$\left| {{T_t}} \right|$$, and for each internal node t in a top-down manner: (11) $\begin{matrix} g (t) = \frac{C (t) - C (T_{t})}{| T_{t} | - 1} \\ α = \min (α, g (t)) \end{matrix}$ $$\begin{array}{c} g(t) = \frac{{C(t) - C\left( {{T_t}} \right)}}{{\left| {{T_t}} \right| - 1}} \\ \alpha = \min (\alpha ,g(t)) \\ \end{array}$$

where T_t denotes the subtree with t as the root node, $C (T_{t})$ $$C\left( {{T_t}} \right)$$ is the error on the training data, and $| T_{t} |$ $$\left| {{T_t}} \right|$$ is the number of leaf nodes of T_t. 4)

Visit internal node t from top to bottom, prune it if there is g(t) = α, and decide its class by majority voting for leaf node t to get tree T.

5)

Set k = k + 1, α_k = α, T_k = T.

6)

If T is not a tree consisting of the root node alone, go back to step 4).

7)

Select the optimal subtree T_α in subsequence T₀, T₁, …, T_n using cross-validation method.

2.4

Educational Data Narratives

The specific methodology for the educational data narrative constructed in this paper is shown in Figure 2. This approach focuses on how to generate visual representations with narrative functions, which consists of two key steps. First, the explicit or implicit learning objectives/intentions in the learning context are converted into data features/rules that can be read by the data processing system. Secondly, data features that can reflect learning intentions are converted into visualization rules/visual elements using data narrative functionality, which are used to highlight specific aspects of the data and guide the user’s attention.

2.5

Design framework

In this section, a framework for designing a visual learning analytics tool based on the “educational data narrative” perspective is constructed as shown in Figure 3. The following section describes in detail the design process of the tool’s visualization content and interaction functions, starting from the two visual learning analysis processes supported by the tool.

2.5.1

Content design

The educational data narrative guides the design goal of communicating insights to users and helping them develop insights into the data.

A prerequisite for achieving this goal is to clarify the user’s goal or intent for applying the tool. Learning goals and pedagogical objectives vary in different learning contexts and need to be defined by the tool designer from both theoretical and practical perspectives.

The tool should focus on both cognitive and social dimensions in its content design to develop insights and presentations of key events in students’ cognitive and social activities.

On the basis of clarifying users’ learning/teaching goals, the tool relies on learning theories, data and data statistics and mining techniques to map and transform the goals into data features and visualization rules in turn.

In this study, the available design principles for information visualization and educational data narratives are sorted out and four core design principles are summarized, which are to exclude the distraction of redundant information, to enhance the perception of important information, to improve the comprehension of multidimensional and complex information, and to draw attention to the information reflected by key data features.

2.5.2

Functional design

According to the visual learning analysis process model, it can be seen that the user’s reasoning decision-making process occurs in the interaction between the user and the tool, including the user’s exploration and validation of the representation results, the knowledge reasoning combining the evidence and individual experience, and the knowledge-based decision-making and feedback in three dimensions.

Therefore, effective support and guidance for the dynamic collaborative decision-making process can be realized by rationalizing interaction functions in these three dimensions.

First, the exploration and validation of the user’s characterization results is usually a cyclical process that is repeated through guidance and feedback.

Designers can consider optimizing the tool’s view operations and navigation operations to help users examine fine-grained details, as well as creating multidimensionally linked views that can provide clearer insights into multidimensional data compared to isolated views.

Second, knowledge-based reasoning is an important part of the process before user-generated decision making, emphasizing the integration of data evidence found by users in the process of exploration and validation and the knowledge and experience of individual users.

This is manifested by providing users with annotations that help them annotate patterns, outliers, and views of interest, thus forming a record of observations, questions, and hypotheses.

Third, the acquired knowledge can further guide the user’s decision-making process and conduct a new round of analysis.

Designers of visual learning analytics tools should fully consider the need for users to regulate the analysis process and design appropriate functionality to support this feedback process.

3

Practical application of data visualization strategies in Civics teaching and learning

3.1

Purpose of Experiment and Subjects

This experiment mainly focuses on the practical application of visualization technology in blended teaching based on the online learning platform, data collection of various educational teaching data generated in the teaching process, and then interpreting these data through visualization technology, converting them into a form that is easy to understand, digging out the educational value behind these teaching data, and summing up the application of visual teaching data in educational teaching value.

In view of the fact that the data analysis of the experiment requires a large amount of data, a large class (the experimental group) consisting of undergraduate classes of two different majors in ideological and political education public courses, with a total of 100 people, was selected for the experiment.

During the experiment, the teaching data generated in the teaching process were collected online and offline at the same time, and the subsequent visualization analysis will be carried out from the overall and local levels to visualize and analyze the results, and intervene and adjust the teaching of the experimental class throughout the experimental process. At the end of the experiment, in order to verify the effectiveness of the application of visualization in teaching, the learning results of the combined class were compared with those of another combined class (control group) taught by the same teacher.

3.2

Experimental Program Design

Teachers teach two classes in the public classroom using an online learning platform. The experiment is set to mobilize students to be actively involved in the teaching process, during which students can send pop-ups for their problems in the teaching process, but not pop-up messages that are not related to the teaching content. After the lesson, students are expected to make personal evaluations or suggestions about the teaching content or individual teachers.

The implementation of this experiment is carried out in the classroom of the public course “Socialist Core Values”, the teaching content is the teaching of a section of knowledge points, in order to ensure that the experiment has a sufficient amount of data, data collection is carried out in two classes. The implementation process consisted of three main stages.

3.2.1

Implementation of teaching and learning

Teaching follows the teacher’s normal process. Prior to the lesson, the teacher pushes the content of the lesson to the students through the online learning platform for students to preview, and during the preview process students can provide feedback on the teacher’s preview material through online learning.

Teachers scan the code to enter the online learning platform to adapt to the teaching environment. Then select the appropriate class and send the QR code of the classroom to the students so that they can enter the classroom. In this process, the teacher can clearly see the students’ sign-in time and name, and the online learning also sends detailed sign-in form data to the teacher. During the teaching process, according to the experimental settings, students can ask pop-up questions about their problems in the learning process, and the teacher can also randomly name students or vote within the class according to the actual situation. After the lesson, a vignette test on the content knowledge points of the section is released. At the end of the lesson, students can evaluate the teaching process, the content, or the individual teacher.

3.2.2

Data acquisition

After the end of the lesson, the data are exported from the online learning platform, including the students’ check-in data, pre-class preview data, pop-up texts generated during the course, the number of pop-ups, the number of random roll calls, and the students’ evaluations after the class, and other data. In the teaching process, data collection is still done online through the online learning platform, and then offline collection methods are combined to supplement and improve the data.

3.2.3

Visualization and analysis of results

Combined with the data collected online and offline, the data will be organized, and then with the help of the tool Excel to data processing and statistics. In the data visualization link, different types of data are visualized, mainly based on specific problem analysis and visualization application needs, using the most appropriate visualization tools. Combined with the results of visualization, analysis of specific problems.

3.3

CART tree generation based on the degree of application of visualization techniques

After the establishment of the CART classification regression analysis software, the CART classification tree can be generated through the degree of teachers in the classroom using visualization techniques with the student test scores of this data model, the specific scope of the analysis is the experimental group of students in school M above and the control group (a total of 200 people), in the ideological and political education course results. Through the Registrar’s Office to obtain the classification of students’ examination results in the course (excellent or ordinary) entered into the behavioral data to establish the specific data table to be calculated CART, this table and then exported to the file, and then through the software to generate CART classification regression tree.

The CART classification regression tree for the degree of application of visualization techniques and grades is shown in Figure 4. The figure shows how many students in both groups received excellent or average grades in the ideology and politics course. From the figure, the following assessment results can be drawn: the teaching method based on visualization technology has a higher rate of students’ excellence, while in traditional teaching, the majority of students’ grades are average, and the number of students with excellence in the experimental group is 51 more than that in the control group. In the visualization classroom, the students’ excellence rate is about 65.67% when the teacher uses more than 5 types of visualization techniques, based on which, one and only one person’s achievement is ordinary when the frequency of using visualization techniques in the classroom is more.

So far, by analyzing the degree of teachers’ application of visualization technology and test scores with the decision tree algorithm, a more intuitive and persuasive analysis of the rules has been made to find out the strategies to improve the teaching effect in ideological and political education, and to reach a reference for assessing the teaching effect of visualization technology in online learning.

3.4

Analysis of the Educational Effectiveness of the Application of Visualization Strategies

3.4.1

Effectiveness of Civic Education Strategies Based on Data Visualization

A semester-long teaching experiment was conducted from September 2023 to November 2023. The effectiveness of the visualization technology teaching strategy was tested by comparing the performance of two groups of students in the Civic and Political Education course before and after the experiment. Before the experiment, the level of Civic and Political Education of the two groups of students was at the same level.

The ideological education test scores of the two groups of students before and after the experiment are shown in Figure 5. From the figure, it can be seen that before the implementation of the visualization of the ideological education strategy, the results of the two groups of students are evenly distributed, most of them are distributed in the range of 60 to 80 points, and only individual students can achieve higher or poorer results. p=0.713>0.05 between the results of the two groups, i.e., there is no significant difference between the two classes before the implementation of the strategy.

With the passage of time, the two groups of students showed significant differences in their performance in Civics education based on different teaching methods. Comparison of the data revealed that the teaching strategy based on data visualization effectively increased the overall student achievement, with the experimental group showing a highly significant difference by 4.71 compared to the control group, and p=0.000. Visualization technology can enhance students’ abilities in information acquisition and case analysis.

3.4.2

Evaluation of the quality of normalized instruction in the application of visualization technology

In this section, through a survey, 50 colleges and universities in a certain city, including the experimental school above, were collected on the application of visualization technology on ideological and political education courses, as well as data related to the evaluation of teaching quality in ideological and political education courses. The degree of application of visualization technology in each school was normalized, and the impact of visualization technology on the teaching quality of ideological and political education courses was analyzed by comprehensively using visualization and statistical analysis software such as Excel and IBM SPSS Statistics, and by using the heat map in the visualization chart.

Figure 6 shows the relationship between the teaching quality of ideological and political education and the degree of application of visualization technology in 50 colleges and universities in a city. From the figure, it can be seen that the distribution of the normalized values of the application of visualization technology in the 50 schools in the city is 0.58 to 0.90, and the corresponding teaching quality of ideological and political education is between 51 and 81, and there is a positive relationship between the teaching quality and the degree of application of visualization technology. Taking School M investigated in this paper as an example, the normalized value of its application degree is about 0.87, and the teaching quality of ideological and political education in this school reaches 80.7. It shows that the use of visualization technology in ideological and political education can improve the quality of teaching by converting abstract data and concepts into intuitive expressions through graphics and images.

3.4.3

Research and Analysis on Satisfaction with Visualization Strategies in Civic and Political Education

In this section, in order to compare the effect of the teaching mode of this online classroom platform combined with visualization technology with the traditional classroom mode, the learning process materials of the students in the experimental group were collected at the end of the semester, and the satisfaction from the interest in learning, classroom efficiency, classroom interactivity, in three dimensions, was compared with the traditional classroom mode by means of questionnaire research. Satisfaction was divided into five levels according to the rating of 1 (very dissatisfied) to 5 (very satisfied) for quantitative comparison.

The results of the satisfaction survey of the experimental group of students with the two teaching modes are shown in Figure 7. As can be seen from the data in the figure, the use of visualization technology improves students’ learning interest, classroom efficiency, and classroom interactivity in the ideological and political classroom, with satisfaction ratings of 4.09, 4.43, and 4.23 in that order, which is an increase of 33.37% to 64.59% over the traditional mode of satisfaction. The results of the research are in line with the expectations of this study, and the online platform combined with visualization technology as an emerging teaching strategy is relatively new to students compared with the traditional teaching mode. The traditional teaching method is limited to the teacher speaking and the students listening, but the new strategy incorporates the benefits of offline and online teaching, which enhances the students’ participation and enthusiasm. The traditional teaching mode lacks supervision in the pre-study before class, which is only a verbal assignment and lacks the check of pre-study effect, while using the online platform, the teacher can release the pre-study tasks in the classroom, and also upload some relevant pre-study materials for students to complete the pre-study before class.

4

Conclusion

This paper designs a visual learning strategy based on ideological and political education, combines with an online learning platform, and researches the effectiveness of the strategy in improving the effect of ideological and political education of college students through teaching comparison experiments and classification regression tree model.

The teaching strategy proposed in this paper can improve students’ performance in ideological and political education, and after the experiment, the average score of the experimental group is 85, while the control group is only 80.29, and there is a significant difference between the two groups. In addition, there is a certain positive relationship between the degree of application of visualization technology and teaching quality. The findings show that visualization technology can improve students’ learning interest, classroom efficiency, and classroom interactivity in ideological and political classrooms, which is 33.37% to 64.59% higher than the traditional mode. By using multiple visualization technologies and increasing the frequency of use, students’ learning performance can be improved.

Język:: Angielski

Częstotliwość wydawania:: 1 razy w roku
Dziedziny czasopisma:: Nauki biologiczne, Nauki biologiczne, inne, Matematyka, Matematyka stosowana, Matematyka ogólna, Fizyka, Fizyka, inne

Kanał RSS czasopisma

Research on the Strategy of Using Data Visualization Technology to Enhance the Effect of Ideological and Political Education of Students in Colleges and Universities

Liya Ji

Data publikacji: 21 mar 2025

Otrzymano: 01 lis 2024

Przyjęty: 17 lut 2025

DOI: https://doi.org/10.2478/amns-2025-0698

Słowa kluczoweVisualization technology, CART classification tree, Teaching experiment, Ideological and political education

© 2025 Liya Ji, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Słowa kluczowe
Visualization technology, CART classification tree, Teaching experiment, Ideological and political education