User-oriented visual data model design method

Abstract

The invention discloses a user-oriented visual data model design method, relates to the technical field of data processing, and solves the problems that the most adaptive standard type is difficult to screen out quickly and accurately according to data characteristics and visual requirements for rich visual types, the visual effect is poor, and the user experience is poor. In order to solve the technical problem that data information cannot be effectively transmitted in the prior art, the method comprises the following steps: analyzing the number of use times of processing visual data in historical data in different time segment intervals, and calculating a fluctuation index to measure the fluctuation condition of data use; the method comprises the following steps: analyzing the occurrence quantity of processing visual data in a time segmentation interval, calculating a rare index to measure the scarcity degree of the data, integrating the two to obtain an analysis index, accurately determining the focus of attention of a user, providing a direction for targeted optimization, collecting the feedback of the user on a standard type layout, analyzing layout defects, and improving the user experience. And making an optimization scheme by referring to success experience of the same type of model.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a user-oriented visual data model design method. Background Art

[0002] With the rapid development of information technology, the amount of data has exploded, and the sources of data have become increasingly extensive, covering corporate databases, Internet public data sets, sensor collected data, etc.

[0003] According to the patent application with publication number CN118467549B, a visual data connection method and device based on a data model are disclosed, which includes: proposing a connection data model description method to describe the characteristics of complex data structures from multiple dimensions, and realizing comprehensive and complete metadata recording and management of complex business data objects; based on the design of the connection data model, proposing a matching visual modeling method to support the rapid and accurate construction of the connection data model; designing a comprehensive connection configuration template to support the automatic generation and execution of connection code based on the template configuration information.

[0004] However, there are still some shortcomings in the existing technology. On the one hand, the original data often has quality problems such as noise, duplication, and missing values, which seriously affect the accuracy and reliability of data analysis. On the other hand, different industries and different users have diverse needs for data visualization. The traditional single and universal data visualization type and layout method are difficult to accurately match the complex and changeable data characteristics and user personalized needs. At the same time, as time goes by, users' use of data visualization models will change, and the existing technology lacks effective tracking and analysis methods for such dynamic changes. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a user-oriented visualization data model design method, which solves the problem that it is difficult to quickly and accurately screen out the most suitable standard type based on data characteristics and visualization requirements when faced with a variety of visualization types, resulting in poor visualization effects and inability to effectively convey data information.

[0006] To achieve the above objectives, the present invention is implemented by the following technical scheme: Based on a user-oriented visual data model design method, the method specifically includes the following steps:

[0007] Obtain visualization data, perform preprocessing on it to obtain preprocessed data, and process it according to visualization requirements to obtain processed visualization data;

[0008] Acquire the data characteristics of the processed visualization data, and filter the visualization types according to the visualization requirements to obtain pre-selected types, and filter the pre-selected types according to the data characteristics to obtain standard types;

[0009] Obtain historical data and divide the time into time intervals. Calculate the mean of the difference in the number of times used within the time intervals and perform normalization to obtain the volatility index. Similarly, analyze the number of occurrences to obtain the rarity index. Sum the two to obtain the analysis index, and compare it with the threshold to select the focus.

[0010] Determine layout shortcomings based on user feedback data of standard types, obtain corresponding optimization solutions, calculate feedback index and matching index of the optimization solutions, sum them up to obtain optimization solution selection values, and generate optimization criteria;

[0011] The layout of the focus is analyzed based on the design method of the optimization standard to generate design information.

[0012] As a further solution of the present invention, the specific method of obtaining the standard type is:

[0013] Analyze and obtain the data characteristics of the obtained processed visualization data, and the data characteristics are data type and data scale, and obtain all visualization types, and at the same time filter the visualization types according to the visualization requirements to obtain pre-selected types, and then filter the pre-selected types according to the data characteristics to obtain standard types;

[0014] The data characteristics corresponding to the pre-selected type are obtained and recorded as the characteristics to be analyzed. At the same time, the characteristics to be analyzed are matched with the data characteristics, and the pre-selected types that match the two are selected and recorded as the standard types.

[0015] As a further solution of the present invention, the specific method of obtaining the volatility index is:

[0016] Obtain all historical data within a time T, and the specific value of time T is set by the operator, obtain the corresponding processing visualization data in the historical data and label it as i, and i=1, 2, j, where j represents the type of processing visualization data;

[0017] The time T is segmented based on time t to obtain the time segment interval, and the specific value of time t is set by the operator. Then the number of times C that the processed visualization data i is used in the time segment interval is obtained. i ;

[0018] In this way, the usage counts of all time segments within time T are obtained, the difference in usage counts between two adjacent time segments is calculated, and the mean of the usage count differences is calculated. The calculated mean is normalized, and the normalized result is recorded as the fluctuation index.

[0019] As a further solution of the present invention, the specific method of obtaining the rare index is:

[0020] The time T is segmented to obtain the time segment intervals, and the number of occurrences corresponding to the processed visualization data i in the time segment intervals within the time T is obtained. Then the number of occurrences is normalized, and the normalized number of occurrences is recorded as the rarity index.

[0021] As a further solution of the present invention, the specific method of screening the focus of attention is:

[0022] The obtained volatility index and rarity index are summed to obtain the analysis index of the processed visualization data. Similarly, the same analysis is performed on all processed visualization data to obtain the corresponding analysis index. The data are sorted from large to small according to the analysis index. The analysis index is then compared with a threshold, and the specific value of the threshold is set by the operator. The processed visualization data with an analysis index greater than the threshold is marked as the focus of attention.

[0023] As a further solution of the present invention, the specific method of generating the optimization criteria is:

[0024] Obtain the standard type and the user feedback data corresponding to the standard type, and obtain the layout shortcomings corresponding to the standard type according to the user feedback data, and then obtain the corresponding optimization plan according to the obtained layout shortcomings, and the optimization plan represents the optimization plan of the same type of model, and is labeled as n, and n=1, 2, m, where m represents the number of optimization plans, and obtain the user feedback data corresponding to the optimization plan, and calculate the feedback index and matching index of the optimization plan according to the user feedback data.

[0025] As a further solution of the present invention, the feedback index method of the calculation optimization scheme is:

[0026] All user feedback data are obtained, and the feedback value corresponding to the optimization scheme n in the user feedback data is obtained, and the mean of all feedback values ​​is calculated. At the same time, the obtained mean is normalized and recorded as the feedback index.

[0027] As a further solution of the present invention, the method of calculating the matching index of the optimization solution is:

[0028] Get the optimization plan, and obtain the standard type by matching analysis from the layout mode. At the same time, get the spatial layout ratio of the optimization plan and the standard type and normalize them. Then analyze the element coverage of the optimization plan and the standard type and normalize them in the same way. Sum the two numerically to get the corresponding matching index.

[0029] As a further solution of the present invention, the specific method of generating the optimization criteria is:

[0030] The feedback index and the matching index are summed to obtain the selection value corresponding to the optimization scheme, and the optimization scheme with the largest selection value is used as the standard to generate the optimization standard.

[0031] The present invention provides a user-oriented visual data model design method. Compared with the prior art, it has the following beneficial effects:

[0032] The present invention analyzes the usage frequency of processed visualization data in different time segments in historical data, calculates the fluctuation index to measure the fluctuation of data usage, analyzes the number of occurrences of processed visualization data in the time segment, calculates the scarcity index to measure the scarcity of data, and combines the two to obtain the analysis index, accurately determine the user focus, and provide direction for targeted optimization.

[0033] The present invention establishes a layout optimization system based on user feedback data, collects user feedback on standard type layouts, analyzes layout shortcomings, draws on the successful experience of similar models to formulate optimization plans, and at the same time calculates the feedback index and matching index, and combines the two to obtain the selection value to determine the most advantageous optimization plan, which significantly improves user satisfaction and operational efficiency. For example, it can effectively increase user residence time and purchase conversion rate on e-commerce platforms. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a diagram of the steps of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Embodiment 1

[0037] See also Figure 1 The present application provides a user-oriented visual data model design method, which specifically includes the following steps:

[0038] Step 1. Obtain visualization data. These data come from a wide range of sources, such as corporate databases, public data sets on the Internet, and data collected by sensors. Taking e-commerce companies as an example, their data may come from order data recorded by the transaction system, browsing and purchasing behavior data collected by the user behavior monitoring system, etc., and data preprocessing is performed to obtain preprocessed data. The data preprocessing operations here include data cleaning, data denoising, duplicate data deletion, and missing value processing. Duplicate data records are identified and deleted through data deduplication algorithms. For missing values, mean filling, median filling, and machine learning-based prediction filling are used for processing. At the same time, the preprocessed data is converted accordingly according to visualization requirements, and then grouped and aggregated to obtain processed visualization data. Taking e-commerce order data as an example, if you want to show the total sales of each category of goods in different months, you can first group the order data by month and product category, and then sum and aggregate the order amount in each group, and finally obtain the total sales data of each category of goods in each month.

[0039] Step 2. After completing the acquisition, preprocessing, grouping and aggregation of the original data, the processed visual data is obtained. At this point, it is necessary to deeply analyze and obtain the key characteristics of the data, mainly covering the two aspects of data type and data scale. Data types can be divided into numeric types (such as integers, floating point numbers, data such as product sales, prices, etc.), character types (such as strings, such as product names, customer names), date and time types (such as order dates, user registration time), etc. The data scale is mainly reflected by the corresponding data capacity. For example, a database table containing millions of transaction records has a relatively large data capacity; while a small data set with only a few hundred records has a small data capacity.

[0040] At the same time, it is necessary to obtain all types that may be applicable to data visualization. Common visualization types include bar charts, line charts, pie charts, scatter plots, maps, etc. Next, these visualization types are preliminarily screened according to specific visualization requirements to obtain pre-selected types. For example, if the visualization requirement is to display the sales share of different product categories, based on this requirement, pie charts, bar charts, etc. can be used as pre-selected types because they are good at showing proportions and comparative relationships; if the requirement is to present the trend of sales of a certain product over time, line charts, area charts, etc. will become pre-selected types.

[0041] Subsequently, the pre-selected types are further screened according to the data characteristics to determine the final standard type. The specific screening method is as follows: for each pre-selected type, the corresponding data characteristics are obtained and recorded as the characteristics to be analyzed. Then, the characteristics to be analyzed are carefully matched with the actual data characteristics of the processed visualization data, and the pre-selected types that match the two are screened out and recorded as the standard types.

[0042] For example, suppose the data to be processed for visualization is the annual sales data of each region of a company. The data types include region (character type) and sales (numeric type). The data size records the sales data of 50 regions, and the data capacity is moderate. The visualization requirement is to compare the sales differences of each region. The pre-selected types obtained through preliminary screening are bar charts and bar charts. For bar charts, the characteristics to be analyzed are suitable for comparing numerical data, and can clearly show the numerical differences between different categories (regions); for bar charts, they are also suitable for comparing numerical data, and when there are many categories, horizontal arrangement is easier to read. Since the data characteristics of the processed visualization data match the characteristics to be analyzed of bar charts and bar charts, both bar charts and bar charts can be used as standard types for the visualization of the data.

[0043] Step 3, obtain standard types and process visualization data, and obtain all historical data within time T. The specific value of time T is set by the operator, and analyze the historical data to determine the user's focus. The specific analysis is determined by analyzing data volatility and data scarcity.

[0044] In order to deeply analyze the usage trend of processing visualization data at different time stages, we first obtain the corresponding processing visualization data in the historical data. These processing visualization data are labeled and recorded as i, where i=1, 2, j, where j represents the total number of types of processing visualization data.

[0045] For example, in an e-commerce sales data visualization project, the processed visualization data may include sales charts by product category, order volume distribution maps by region, user purchase frequency line charts, etc. In this case, j is the number of these different types of visualization data.

[0046] Next, the total time range T is segmented based on time t, thereby obtaining a series of time segment intervals. The specific value of time t is set by the operator according to actual analysis requirements.

[0047] For example, if the time range T of interest is one year (from January 1 to December 31), and the operator analyzes at intervals of one month, then t is set to one month, and the time segment intervals are January 1-January 31 and February 1-February 28 respectively.

[0048] Then, for each type of processed visualization data labeled i, count its usage times C in each time segment interval i For example, for the sales chart by product category (assuming that its label i=1, in the time segment from January 1 to January 31, it was viewed 15 times by the market analyst, then in this interval C 1 =15.

[0049] And so on, the number of times each type of processed visualization data is used in all time segment intervals within time T is obtained.

[0050] After obtaining the usage count of each time segment, calculate the difference between the usage counts of two adjacent time segment. Assume that for the processed visualization data labeled i, the usage count in the kth time segment is C i，k , the number of times used in the k+1th time segment is C i，k+1 , then the difference in the number of times used in two adjacent time segments is =C i，k+1 -C i，k .

[0051] For example, for the above sales chart i=1, the number of times C is used from January 1st to January 31st 1，1 =15, usage times C from February 1 to February 28 1，2 =20, then the difference in the number of times adjacent intervals are used is 20-15=5.

[0052] Calculate the mean of the difference in the number of times used in all adjacent time segment intervals. For the processed visualization data labeled i, assume that there are a total of n time segment intervals ( ), then the mean of the difference in the number of uses .

[0053] Continuing with the sales chart above, assuming there are 12 months in a year, n=12, the difference in usage times in 11 adjacent months is calculated as , , …, , add these differences and divide by 11 to get the mean of the usage times difference .

[0054] Finally, the calculated mean is normalized. Normalization allows the fluctuations of different types of processed visualization data to be compared on the same scale. There are many commonly used normalization methods, such as maximum-minimum normalization.

[0055] Assume that the maximum value among the mean differences in the number of times all processed visualization data are used is , the minimum value is For the processed visualization data labeled i, the normalized result is the volatility index Volatility Index FI i The value range is between 0 and 1. The larger the value, the more drastic the fluctuation of the number of times this type of processing visualization data is used within time T; the smaller the value, the more stable its use.

[0056] The specific method of analyzing data scarcity is to segment the time T to obtain the time segment interval, and obtain the number of occurrences corresponding to the processed visualization data i in the time segment interval within the time T, and then normalize the number of occurrences. The normalization method here is the same as the normalization method of data volatility, and the normalized number of occurrences is recorded as the scarcity index;

[0057] The obtained volatility index and rarity index are summed to obtain the analysis index of the processed visualization data. Similarly, the same analysis is performed on all processed visualization data to obtain the corresponding analysis index. The data are sorted from large to small according to the analysis index. The analysis index is then compared with a threshold, and the specific value of the threshold is set by the operator. The processed visualization data with an analysis index greater than the threshold is marked as the focus of attention.

[0058] Step 4. First, determine the standard type of layout. Standard types usually come from widely recognized best practices, established design specifications or mature design patterns in the industry. For example, in the layout design of the product display page of an e-commerce platform, standard types may include waterfall layout (products are displayed in a continuous vertical arrangement without paging), grid layout (products are displayed in neat rows and columns), etc. After determining the standard type, collect corresponding user feedback data. This can be achieved in many ways, such as setting up a user satisfaction questionnaire on the platform to ask users about the ease of use, visual effects, and convenience of finding products in the current layout; or through user behavior analysis tools, record the user's operation trajectory on the page, stay time and other data, so as to gain insight into the user's actual experience of the layout. Based on the collected user feedback data, deeply analyze the shortcomings of the standard type layout. For example, for the waterfall layout, user feedback data shows that some users spend a long time looking for specific products, which indicates that the layout is insufficient in the product positioning function; for the grid layout, it may be found that users frequently experience visual fatigue during browsing, indicating that the visual design or element arrangement of the layout needs to be improved.

[0059] Find corresponding optimization solutions for the identified layout shortcomings. These optimization solutions usually draw on the successful experience of similar models. Taking the e-commerce platform as an example, if the product positioning problem of the waterfall layout is prominent, the optimization solution may be to add a product classification navigation in the sidebar of the page, or introduce an intelligent search and filtering function; for the visual fatigue problem of the grid layout, the optimization solution may be to adjust the display style of the product pictures and increase the blank space between elements to optimize visual comfort. These optimization solutions are labeled as n, and n=1, 2, m, where m represents the number of optimization solutions.

[0060] For each optimization solution, relevant user feedback data is also collected. Based on this, two key indicators are calculated: feedback index and matching index.

[0061] Feedback index calculation: Comprehensively collect all user feedback data related to optimization plan n and extract the feedback value. The feedback value can be the user's rating in the questionnaire (such as 1-5 points), or the positive or negative sentiment score extracted from the user comments through sentiment analysis. Calculate the mean of these feedback values ​​and then normalize them. Assume that the user ratings received by optimization plan 1 (adding a sidebar navigation in the waterfall layout) are 3, 4, 5, 4, and 3, respectively, and its mean is (3+4+5+4+3)÷5=3.8. If the maximum value of the mean feedback values ​​of all optimization plans is 4.5 and the minimum value is 2.5, then the feedback index of the optimization plan is (3.8−2.5)÷(4.5−2.5)=0.65.

[0062] Matching index calculation: Taking optimization scheme 1 as an example, firstly, from the perspective of layout mode, a matching analysis is conducted with the standard type (waterfall flow layout), the similarity between the optimization scheme and the standard type in spatial layout is determined, and the spatial layout ratio is calculated. Assuming that the product display area in the standard type accounts for 70% of the total page area, and the ratio in optimization scheme 1 is 68%, the spatial layout ratio is 68%÷70%≈0.97. All relevant spatial layout ratios are normalized. If the maximum value is 1.1 and the minimum value is 0.8, the spatial layout ratio of optimization scheme 1 is normalized to 0.97÷1.1≈0.88. At the same time, the element coverage of the optimization scheme and the standard type is analyzed. For example, the product image element in the standard type should cover 40% of the total page area, and the actual coverage area of ​​the product image in optimization scheme 1 is 38% of the total page area, then the element coverage is 38%÷40%=0.95. The element coverage is normalized. If the maximum value of all element coverage is 1.05 and the minimum value is 0.9, the element coverage of optimization scheme 1 is normalized to 0.95÷1.05≈0.90. Adding the normalized spatial layout ratio and element coverage, the matching index of optimization scheme 1 is 0.88+0.90=1.78.

[0063] Add the feedback index and matching index of each optimization solution to get the corresponding selection value. Continuing with the above optimization solution 1 as an example, the selection value is 0.65+1.78=2.43. By comparing the selection values ​​of all optimization solutions, the optimization solution with the largest selection value is the most advantageous solution. In the case of the e-commerce platform, if the selection values ​​of other optimization solutions are all less than 2.43, then optimization solution 1 (adding a sidebar navigation in the waterfall layout) will be determined as the final optimization standard.

[0064] By driving the generation of optimization solutions through user feedback data, we can accurately solve the pain points encountered by users in the process of using the layout, and significantly improve user satisfaction and operational efficiency. For example, the optimized layout of the e-commerce platform can enable users to find the required products faster, reduce browsing fatigue, and thus increase the user's stay time on the platform and the purchase conversion rate.

[0065] Step 5. Obtain the optimization standard and focus, and design the focus based on the design method of the optimization standard to generate corresponding design information.

[0066] Some of the data in the above formulas are calculated by taking their numerical values, and are not substituted into parameter units for calculation. At the same time, the contents not described in detail in this specification belong to the existing technologies known to those skilled in the art.

[0067] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A user-oriented visual data model design method, characterized in that: The method specifically comprises the following steps: Obtain visualization data, perform preprocessing on it to obtain preprocessed data, and process it according to visualization requirements to obtain processed visualization data; Acquire the data characteristics of the processed visualization data, and filter the visualization types according to the visualization requirements to obtain pre-selected types, and filter the pre-selected types according to the data characteristics to obtain standard types; Obtain historical data and divide the time into time intervals. Calculate the mean of the difference in the number of times used within the time intervals and perform normalization to obtain the volatility index. Similarly, analyze the number of occurrences to obtain the rarity index. Sum the two to obtain the analysis index, and compare it with the threshold to select the focus. Determine layout shortcomings based on user feedback data of standard types, obtain corresponding optimization solutions, calculate feedback index and matching index of the optimization solutions, sum them up to obtain optimization solution selection values, and generate optimization criteria; The layout of the focus is analyzed based on the design method of the optimization standard to generate design information.

2. The user-oriented visual data model design method according to claim 1, characterized in that: The specific method of obtaining the standard type is: Analyze and obtain the data characteristics of the obtained processed visualization data, and the data characteristics are data type and data scale, and obtain all visualization types, and at the same time filter the visualization types according to the visualization requirements to obtain pre-selected types, and then filter the pre-selected types according to the data characteristics to obtain standard types; The data characteristics corresponding to the pre-selected type are obtained and recorded as the characteristics to be analyzed. At the same time, the characteristics to be analyzed are matched with the data characteristics, and the pre-selected types that match the two are selected and recorded as the standard types.

3. The user-oriented visual data model design method according to claim 1, characterized in that: The specific method of obtaining the volatility index is: Obtain all historical data within a time T, and the specific value of time T is set by the operator, obtain the corresponding processing visualization data in the historical data and label it as i, and i=1, 2, j, where j represents the type of processing visualization data; The time T is segmented based on time t to obtain the time segment interval, and the specific value of time t is set by the operator. Then the number of times C that the processed visualization data i is used in the time segment interval is obtained. i ; In this way, the usage counts of all time segments within time T are obtained, the difference in usage counts between two adjacent time segments is calculated, and the mean of the usage count differences is calculated. The calculated mean is normalized, and the normalized result is recorded as the fluctuation index.

4. The user-oriented visual data model design method according to claim 1, characterized in that: The specific method of obtaining the rare index is: The time T is segmented to obtain the time segment intervals, and the number of occurrences corresponding to the processed visualization data i in the time segment intervals within the time T is obtained. Then the number of occurrences is normalized, and the normalized number of occurrences is recorded as the rarity index.

5. The user-oriented visual data model design method according to claim 1, characterized in that: The specific method of screening the focus is as follows: The obtained volatility index and rarity index are summed to obtain the analysis index of the processed visualization data. Similarly, the same analysis is performed on all processed visualization data to obtain the corresponding analysis index. The data are sorted from large to small according to the analysis index. The analysis index is then compared with a threshold, and the specific value of the threshold is set by the operator. The processed visualization data with an analysis index greater than the threshold is marked as the focus of attention.

6. The user-oriented visual data model design method according to claim 1, characterized in that: The specific method of generating the optimization criteria is: Obtain the standard type and the user feedback data corresponding to the standard type, and obtain the layout shortcomings corresponding to the standard type according to the user feedback data, and then obtain the corresponding optimization plan according to the obtained layout shortcomings, and the optimization plan represents the optimization plan of the same type of model, and is labeled as n, and n=1, 2, m, where m represents the number of optimization plans, and obtain the user feedback data corresponding to the optimization plan, and calculate the feedback index and matching index of the optimization plan according to the user feedback data.

7. The user-oriented visual data model design method according to claim 6, characterized in that: The feedback index method of the calculation optimization scheme is: All user feedback data are obtained, and the feedback value corresponding to the optimization scheme n in the user feedback data is obtained, and the mean of all feedback values ​​is calculated. At the same time, the obtained mean is normalized and recorded as the feedback index.

8. The user-oriented visual data model design method according to claim 6, characterized in that: The method of calculating the matching index of the optimization solution is: Get the optimization plan, and obtain the standard type by matching analysis from the layout mode. At the same time, get the spatial layout ratio of the optimization plan and the standard type and normalize them. Then analyze the element coverage of the optimization plan and the standard type and normalize them in the same way. Sum the two numerically to get the corresponding matching index.

9. The user-oriented visual data model design method according to claim 1, characterized in that: The specific method of generating the optimization criteria is: The feedback index and the matching index are summed to obtain the selection value corresponding to the optimization scheme, and the optimization scheme with the largest selection value is used as the standard to generate the optimization standard.

Images