CN106296350A - A kind of visual analyzing city public bicycle system borrows the method for also pattern - Google Patents

Abstract

The invention discloses a method for visually analyzing the borrowing and returning mode of an urban public bicycle system. The steps of the invention are as follows: 1. Collect public bicycle data, and preprocess the data; 2. Design a geographical view based on the spatial perspective, visually display the geographical location distribution of the site, and provide a spatial filtering function at the same time to help analysts select sites interactively Or a collection of sites; 3. Design a single-site borrowing and repaying time-domain thermal view based on a time perspective; use a visual coding method similar to a table to display the changes and differences in the amount of borrowing and repaying a site in different time periods; 4. Based on a spatial perspective Design the site loan-return relationship view to display the many-to-many site loan-return relationship; 5. Design a multi-attribute view. The invention can effectively improve traffic managers' awareness of the operation of the public bicycle system, improve data analysis efficiency, and provide auxiliary decision-making for site management and vehicle scheduling.

Description

A Method for Visual Analysis of Borrowing and Returning Patterns of Urban Public Bicycle System

technical field

The invention belongs to the field of information technology, and in particular relates to a method for visually analyzing the loan-return mode of an urban public bicycle system.

Background technique

The urban public bicycle system provides a shared bicycle rental service, which is a new and green public transportation mode. In the past 10 years, major cities in the world, such as Paris, New York, Barcelona, etc., have established public bicycle systems.

In a city with a public bicycle system, there are public bicycle stations all over the place. There are multiple immobilizers in each station for parking volume. The user borrows the car from the station closer to the departure point through the IC card, and returns the car to the station closer to the destination. Using advanced network technology, vehicle leasing information can be tracked and stored in the database. This information contains the digital footprint of users moving in urban space, which is an important basis for understanding system operation and assisting traffic decision-making.

When users use the system, they can freely select the borrowing and returning route and time based on different travel needs every day, resulting in a high degree of variability in the vehicle borrowing and returning trajectory. Traffic management personnel usually do not have professional data analysis and database operation knowledge, and cannot intuitively understand the borrowing and returning mode of vehicles in the system for management and decision-making. The method of visual analysis can mine the laws hidden in the public bicycle data collection. Existing visual analysis methods use simple histograms, line graphs, etc. to display the spatio-temporal attributes of data, not only do not design suitable visual representations for the inherent characteristics of the data, but also lack of multiple factors that affect vehicle borrowing and returning (such as weather conditions, Comprehensive analysis of holidays, etc.).

Therefore, it is necessary to design a visual view combining the perspectives of time, space, and multi-dimensional attributes, allowing users to intuitively analyze data rules through easy-to-understand graphical representations, and support user-driven interactive exploration of the borrowing and returning modes of urban public bicycle systems, thereby helping Traffic staff can better manage the system and dispatch vehicles, relieve urban traffic pressure, and can also provide suggestions for citizens to travel by bicycle.

Contents of the invention

The purpose of the present invention is to propose a method for visually analyzing the borrowing and returning mode of the urban public bicycle system for dynamic and multivariate public bicycle data sets. The present invention designs multiple visual views to show the characteristics of data sets in terms of time, space and multi-dimensional attributes, helps management departments understand the main functions of site operations, digs out the mainstream direction of vehicle movement within a certain period of time, and analyzes the impact of various factors on the number of loans and repayments Impact.

The technical scheme that the present invention adopts comprises the steps as follows:

Step 1: Collect public bicycle data and preprocess the data.

Step 2: Based on the spatial perspective, design the geographical view, and visually display the geographical distribution of the sites through the geographical view, and provide the spatial filtering function to help analysts interactively select sites or site collections.

Step 3: Based on the time perspective, design a single-site borrowing and repaying time-domain thermal view. Using a visual coding method similar to a table, it can display the changes and differences in the borrowing and repayment amount of a certain station in different time periods, and discover the stations and time periods of long-term vehicle shortages, so as to better assist vehicle scheduling.

Step 4: Based on the spatial perspective, design the site loan-return association view to show the many-to-many site loan-return relationship.

Step 5: Design a multi-attribute view, mainly to analyze the impact on the number of borrowed and returned vehicles under the influence of multiple factors such as different weather conditions and date attributes.

Said step 1 includes:

Step 1.1: Obtain the bicycle rental data collection for the past 3 months. Among them, the bicycle rental data table stores the relevant information of all users renting a car. A rental record is as follows:

hire_r＝[uID, bikeID, cardNo, leaseStat, leaseTime, returnStat, returnTime]

Respectively represent the user ID, vehicle ID, user card number, car rental station, car rental time, car return station and car return time.

The station information table stores information related to bicycle stations, and a station record is as follows:

statInfo_r=[statID, statName, statAddr, lng, lat, serviceTime]

Respectively represent the site ID, site name, site address, longitude and latitude, and service time.

Step 1.2: Preprocess the data and delete useless rental records.

Useless rental records include 3 categories: (1) The station where the vehicle is returned is null (returnStat=null), which indicates that the vehicle may be lost. (2) The borrowing time is longer than the returning time (leaseTime>returnTime), which indicates that the data record is wrong. (3) The time interval between borrowing the car and returning the car is less than 3 minutes (returnTime-leaseTime<3min), which indicates that the user did not actually borrow the car. It may be that the car was unable to ride due to a fault with the car, so the car was returned quickly.

Said step 2 includes:

Step 2.1: Using the Baidu map interface, based on the longitude and latitude of the site, represent the site with an icon on the map. When the user clicks on the site icon, information related to the site pops up.

Step 2.2: Provide an interface for site or site collection selection. Support to select single or multiple sites by clicking on the map; provide an area lock tool, allowing analysts to draw a circular area on the map to get all the sites in the area.

Said step 3 includes:

Step 3.1: Design the sub-view of borrowing and repaying heat. The number of table rows is the number of dates to be analyzed. The number of table columns is 16, corresponding to the opening hours of public bicycles from 6 am to 9 pm. Each row represents a day selected by the user. The color of the row label is used to distinguish different categories of dates, black for weekdays, blue for weekends, and red for small holidays. Each cell displays the amount of borrowed cars and returned cars for a certain hour on a certain day at the same time. The left is the amount of borrowed cars, and the right is the amount of returned cars. The color of the cell is proportional to the loan amount. The darker the orange, the larger the value. Find the maximum and minimum values by taking the borrowed and returned vehicles for all eligible hours, and then normalize each value and map it to a color scale.

Step 3.2: Design the sub-view of the loan difference heat. At this time, each cell in the table represents the difference between the amount of borrowed cars and the amount of returned cars in a certain hour on a certain day, and the difference is mapped to a red-white-blue color scale. The redder the cell color, it means that the amount of borrowed cars is far greater than the amount of returned cars, that is, there are many empty parking spaces in the site, and users may have no cars to borrow. The bluer the cell color, the more cars are returned than the borrowed cars, that is, the site is full of cars, and the user may not be able to return the car. When the amount of borrowed cars and returned cars tends to be equal, the color of the cell is white.

Said step 4 includes:

Step 4.1: Design the sub-view of the vehicle space direction to show the geographical space connection of the stations. When the user selects a central site, the view calculates the top N sites most closely associated with the central site based on the rental amount within the specified time period, and connects the related sites with arcs on the map. At the center of each arc is a small arrow indicating the direction in which to borrow or return. The thickness and transparency of the arc is related to the corresponding borrow/repay amount. The thicker and more opaque the line, the greater the amount. When the user selects multiple sites, the relationship between multiple sites is also drawn on the map with an arc based on the loan amount.

Step 4.2: Design a multi-site traffic correlation sub-view, and use the chord diagram to visually encode the borrowing and repaying quantities between multiple sites. An arc corresponds to a station. The length of the arc is proportional to the sum of loans and repayments of the site during the specified time period. A chord encodes the difference in debit and repayment amounts between two sites. Suppose a chord starts from arc A and goes to arc B. The length of the chord on arc A represents the number of bicycles borrowed from station A and returned to station B. Correspondingly, the length of the chord on arc B represents the number of bicycles borrowed from station B and returned to station A. If the lengths of a chord on the two end arcs are very different, it indicates that the two-way traffic between the two sites is very different. For a particular site, if the amount of borrowed cars is large and the amount of returned cars is small, it means that many people start from that site. Conversely, if many people return cars to this station, but few people borrow cars from this station, it shows that this is a destination station. When the analyst moves the mouse over an arc, only the chords associated with that arc are displayed.

Said step 5 includes:

Step 5.1: Generate a multi-factor data table to store the date attributes and weather conditions of each day in the given analysis time period. The date attribute (is_holiday) has three attribute values: working day, weekend, and small holiday. Grab the weather conditions from the Internet and store them in the data table. Including average temperature attribute (avgTemp), weather attribute (weather) and wind speed attribute (wind). The weather attribute has seven attribute values: sunny, cloudy, showers, light rain, moderate rain, heavy rain, and snow. The wind speed attribute has four attribute values: the wind force is less than 3, the wind force is 3-4, the wind force is 4-5, and the wind force is greater than 5. Analyze the impact of the above attributes on the number of bikes borrowed (bikeNum), and consider the number of bikes borrowed as an attribute. These attributes can be divided into two categories: numerical attributes (avgTemp, bikeNum) and categorical attributes (is_holiday, weather, wind). The attribute values of categorical attributes are discrete and only include some specific values, while the attribute values of numeric attributes are continuous.

Step 5.2: Design a new line- and set-based parallel coordinates component that simultaneously exhibits the characteristics of multivariate datasets with both categorical and numerical attributes.

Step 5.2.1: Based on the attribute characteristics, draw the coordinate axes, corresponding to five attributes from left to right: avgTemp, bikeNum, weather, isHoliday, wind. Draw five axes parallel to each other and perpendicular to the horizontal plane. The first two axes represent numerical attributes, represented by a straight line, with corresponding coordinates on the straight line, and connected by straight lines to represent the relationship between the coordinate axes. The last three axes represent category attributes, which are represented by a rectangle, and each attribute value occupies a small section of the rectangle, which is called an axis column. The color of the axis column is used to distinguish different attribute values, and the number of axis columns is the number of values of all attribute values. An axis column is further divided into sub-axis columns according to the proportion of a site's current property value. Connect the two sub-axle columns with a quadrilateral.

For two parallel category attribute axes, in order to calculate the height of the axis column and the width of the quadrilateral. First, for a given site statID, retrieve and generate records containing multiple factors that affect the amount of car rental, where _date represents the date of a certain day:

multiFac_rec＝[statID,date,avgTemp,bikeNum,weather,isHoliday,wind]

{multiFac_rec} _statID indicates the multi-factor record set of statID site. Retrieve the number f _k,i,j of eligible valid data items from {multiFac_rec} _statID , where k corresponds to the site ID, i is an attribute value on the left axis, and j is an attribute on the right axis value. Suppose the left axis represents weather and the right axis represents isHoliday. When attribute value (i) = "sunny", attribute value (j) = "weekday", then f _{k, i, j} represent in {multiFac_rec} _statID , satisfy statID = k, weather = "sunny", isHoliday = Number of data items for "weekdays". The width of the quadrilateral connecting the two sub-columns is determined by freq _k,i,j , The length of each axis column on the left axis is proportional to sum_lAxis_freq, representing the frequency of occurrence of each attribute value, Similarly, the length of each column on the right axis is proportional to sum_rAxis_freq, A legend for station names is shown above, and a legend for axis column colors is shown at the bottom.

When two adjacent axes respectively represent a numerical attribute and a category attribute, the lines starting from the value attribute converge to the center point of the category attribute axis column.

Step 5.2.2: Since directly displaying all the data items seems messy, in order to more clearly mine the relationship between multiple influencing factors, provide a way to interact with components to help analysts filter data. When the mouse moves over a line connecting two numerical attributes, the line will be strengthened, and a prompt box will pop up for the values of all attributes associated with the line. When the analyst selects a quadrilateral, all relevant connections are displayed and irrelevant connections are hidden.

The four views in Step 2, Step 3, Step 4, and Step 5 are interrelated. When the time and space filter conditions are changed, the related views are automatically updated. If the user wants to mine the characteristics of a site, the single-site lending and returning time-domain thermal view, the vehicle spatial direction sub-view in the site borrowing and returning association view, and the multi-attribute view are updated. At this time, the number of analysis sites in the multi-attribute view is 1. When the user wishes to compare multiple sites, the two subgraphs and the multi-attribute view in the site loan-return association view are updated.

Compared with the prior art, the present invention has advantages and effects as follows:

The characteristic and innovation of the method of the present invention is to propose a new method for detecting the dynamic borrowing and returning mode of public bicycles, and through the perspective of comprehensive space, time and multi-dimensional attributes, a visual representation suitable for the inherent characteristics of the data is designed to present the data hidden in the data. characteristics; a new visual representation method "parallel coordinates based on lines and sets" is proposed, which simultaneously displays the characteristics of multivariate data sets with category attributes and numerical attributes, and supports analysts to interactively mine the impact of different heterogeneous factors on car rental Quantitative impact. The invention can effectively improve traffic management personnel's awareness of the operation of the public bicycle system, improve the efficiency of data analysis, and provide auxiliary decision-making for site management and vehicle scheduling.

Description of drawings

The present invention will be further described below in conjunction with accompanying drawing.

Fig. 1 is a flow chart of the visualization analysis method of the present invention.

Figure 2 is a diagram of the process of analyzing a single site "Qiuzhi Primary School".

FIG. 3 is a process diagram of analyzing the mainstream direction of vehicle movement within a specific time period.

Figure 4 is a process diagram for analyzing the impact of multiple sites and multiple factors on the rental amount.

detailed description

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

FIG. 1 is a flow chart of the visualization analysis method of the present invention. The user first selects the analysis time period and then specifies whether to analyze a single site or a collection of sites. When the user selects a site in the geographical view, the time-domain thermal view of single-site borrowing and returning, the sub-view of vehicle space direction in the site borrowing and returning association view, and the multi-attribute view are updated synchronously, and are comprehensively used to analyze the characteristics of the site. When the user selects multiple sites, the two sub-views and the multi-attribute view of the site borrowing and returning association view are updated synchronously, showing the associations and differences between multiple sites.

Figure 2 shows the analysis process diagram of a single site "Qiuzhi Primary School". Figure 2(a) is a sub-view of the borrowing and returning heat of the site within a week. It can be seen from the figure that the site has a large amount of borrowing/returning vehicles during the morning and evening peaks of weekdays. The specific time is concentrated from 7:00 am to 8:00 am and from 4:00 pm to 5:00 pm, which corresponds to the time for students to go to school and leave school . Figure 2(b) is a sub-view of the borrowing and repaying differences in the corresponding time period. It can be seen that during the morning rush hour, there are more dark red cells, indicating that many people borrow cars from here. During the evening rush hour, there are more dark blue cells, indicating that many people return their cars here in the evening. Figure 2(c) shows a sub-view of the vehicle's spatial orientation centered on the site (indicated by the five-pointed star icon). By clicking the icon of the associated site to view the specific site information, it is found that the sites closely related to the site mainly include bus stations, residential areas and factories. Based on the above analysis results, it can be concluded that this station is mainly used for commuting, and it is recommended that management personnel pay attention to the removal and replenishment of vehicles in the morning and evening peak periods of the station.

Figure 3 shows the process diagram of analyzing the mainstream direction of vehicle movement in a specific time period. In order to find out the mainstream direction of citizens browsing the West Lake Scenic Area during the holidays, the analyst first selects multiple sites around the West Lake Scenic Area on the geographic view, and selects April 5, a day during the small long holiday, to view the multi-site traffic correlation subview ( Figure 3(a)). When the mouse moves to the arc representing the site of "Children's Palace", relevant information about borrowing a car from this site is displayed. From the length of the arc, it can be seen that the Children's Palace has the largest number of borrowed and returned cars. Borrowing a car from "Children's Palace", most of the cars have returned to "Pinghu Qiuyue", followed by "Quyuan Fenghebei" and "Hangzhou Flower Garden", and from the length of the two ends of the string, the traffic flow difference between these stations is obvious . This shows that "Children's Palace" is a big starting point. Then move the mouse to the arc representing the "Pinghu Qiuyue" station, and find that most of the cars borrowed from this station are returned to "Su Causeway South Exit". Continue to move the mouse to the arc of "South Exit of Su Causeway", and find that the largest bus return site is "Changqiao". From this, we can deduce a mainstream browsing path by the West Lake: "Children's Palace" → "Pinghu Qiuyue" → "Su Causeway South Exit" → "Long Bridge". Figure 3(b) shows the geographical location of the route, which is the direction from the west north line to the south line.

Figure 4 shows the process diagram of the impact of multiple sites and multiple factors on the rental volume. The analyst first selected four sites: "Jinqiu Building", "Pinghu Qiuyue", "Children's Palace", and "Qiuzhi Primary School". Figure 4(a) is an overview of the generated parallel coordinate components based on lines and sets, from which it is found that the records with the highest rental volume are generated by the "Children's Palace" site on weekends and holidays. "Golden Autumn Building" has a relatively high average rental volume, which can be seen from the relatively high position of the green line in the bikeNum axis. "Qiuzhi Primary School" borrows a relatively small amount of cars. When the mouse moves over a line, you can view the values of related properties. Figure 4(b) shows the interactive results of the Pinghu Qiuyue site when the weather is "cloudy" and isHoliday is "weekend". It is found that users borrow cars to travel more on weekends when the weather is better.

Claims (7)

1. A method for visually analyzing the borrowing and returning modes of urban public bicycle systems, characterized in that multiple visual views are designed to show the characteristics of data sets in time, space and multidimensional attributes, thereby mining the mainstream direction of vehicle movement within a certain period of time, To analyze the influence of various factors on the loan and repayment quantity, the specific steps are as follows: Step 1: Collect public bicycle data and preprocess the data; Step 2: Based on the spatial perspective, design a geographic view, visually display the geographical distribution of sites through the geographic view, and provide spatial filtering functions to help analysts interactively select sites or site collections; Step 3: Based on the perspective of time, design a time-domain thermal view of borrowing and repaying a single station; use a visual coding method similar to a table to display the changes and differences in the amount of borrowing and repaying a certain station in different time periods, and find out the stations and locations with long-term vehicle shortages Time period, to better assist vehicle scheduling; Step 4: Based on the spatial perspective, design the site loan-return relationship view to display the many-to-many site loan-return relationship; Step 5: Design a multi-attribute view, mainly to analyze the impact on the number of borrowed and returned vehicles under the influence of multiple factors such as different weather conditions and date attributes. 2. a kind of method of visual analysis urban public bicycle system borrowing and returning mode according to claim 1, it is characterized in that described step 1 step is as follows: Step 1.1: Obtain the bicycle rental data collection for the past 3 months; the bicycle rental data table stores the relevant information of all users who rent a car; a rental record is as follows: hire_r＝[uID, bikeID, cardNo, leaseStat, leaseTime, returnStat, returnTime] Respectively represent the user ID, vehicle ID, user card number, car rental station, car rental time, car return station and car return time; The station information table stores information related to bicycle stations, and a station record is as follows: statInfo_r=[statID, statName, statAddr, lng, lat, serviceTime] Respectively represent the site ID, site name, site address, longitude and latitude, and service time; Step 1.2: Preprocess the data and delete useless rental records; Useless rental records include 3 categories: (1) The return site is null (returnStat=null), which indicates that the vehicle may be lost; (2) The rental time is greater than the return time (leaseTime>returnTime), which indicates that the data record Incorrect; (3) The time interval between borrowing the car and returning the car is less than 3 minutes (returnTime-leaseTime<3min), which indicates that the user did not really borrow the car, and it may be that the car was unable to ride due to a fault with the car, so the car was returned quickly. return. 3. A kind of method of visual analysis urban public bicycle system borrowing and returning mode according to claim 2, it is characterized in that said step 2 comprises: Step 2.1: Use the Baidu map interface, based on the longitude and latitude of the site, use an icon to represent the site on the map; when the user clicks the site icon, pop-up site-related information; Step 2.2: Provide an interface for selecting a station or a collection of stations; support clicking on the map to select a single or multiple stations; provide an area lock tool to allow analysts to draw a circular area on the map to get all the stations in the area. 4. a kind of method of visual analysis urban public bicycle system borrowing and returning mode according to claim 3, it is characterized in that described step 3 comprises: Step 3.1: Design the sub-view of loan and repayment heat; the number of table rows is the number of dates to be analyzed; the number of table columns is 16 columns, corresponding to the opening hours of public bicycles from 6:00 am to 9:00 pm; each row represents The day selected by the user; the color of the row label is used to distinguish different types of dates, black means weekdays, blue means weekends, and red means small holidays; each cell simultaneously displays the amount of borrowed cars and returned cars for a certain hour on a certain day The amount of borrowed cars is on the left, and the amount of returned cars is on the right; the color of the cell is proportional to the amount borrowed and returned; the darker the orange, the larger the value; by obtaining the amount of borrowed cars and returned cars in all hours that meet the conditions , find the maximum and minimum values, and then normalize each value and map it to the color scale; Step 3.2: Design the sub-view of borrowing and returning difference heat; at this time, each cell in the table represents the difference between the amount of borrowed cars and the amount of returned cars in a certain hour on a certain day, and the difference is mapped to a red-white-blue Under the color scale; the redder the color of the cell, the amount of borrowed cars is far greater than the amount of returned cars, that is, there are many empty parking spaces in the site, and the user may not have a car to borrow; the bluer the color of the cell, the amount of returned cars is far greater than the amount borrowed. Vehicle volume, that is, the site is full of vehicles, and the user may not be able to return the vehicle; when the amount of borrowed vehicles and the amount of returned vehicles tend to be equal, the color of the cell is white. 5. A kind of method of visual analysis urban public bicycle system borrowing and returning mode according to claim 4, it is characterized in that said step 4 comprises: Step 4.1: Design the sub-view of the vehicle space direction to show the connection of the sites in geographical space; when the user selects a central site, this view calculates the top N sites most closely related to the central site based on the rental amount within the specified time period , and use arcs to connect related stations on the map; there is a small arrow in the center of each arc, indicating the direction of borrowing and returning; the thickness and transparency of the arc are related to the corresponding borrowing/repaying amount; the thicker the line, the more opaque, Indicates that the quantity is larger; when the user selects multiple sites, the relationship between multiple sites is also drawn on the map with an arc based on the loan amount; Step 4.2: Design a multi-site traffic correlation sub-view, using a chord diagram to visually encode the amount of borrowing and repayment between multiple sites; an arc corresponds to a site; the length of the arc is proportional to the sum of the amount of borrowing and repaying the site within a specified time period; an arc The chord encodes the difference in borrowing and repayment between two stations; if the length of a chord on the arcs at both ends is very different, it means that the two-way traffic between the two stations is very different; for a specific station, if borrowing a car If the amount of vehicles is large, but the amount of vehicles returned is small, it indicates that many people start from that site; on the contrary, if many people return vehicles to this site, but few people borrow cars from this site, it indicates that this is a destination site; when analyzing Or when the mouse moves to an arc, only the chords related to the arc are displayed. 6. A kind of method of visual analysis urban public bicycle system borrowing mode according to claim 5, it is characterized in that described step 5 comprises: Step 5.1: Generate a multi-influencing factor data table to store the date attribute and weather conditions of each day within a given analysis period; the date attribute (is_holiday) has three attribute values: weekdays, weekends, and small holidays; grab from the Internet Take the weather conditions and store them in the data table; including the average temperature attribute (avgTemp), weather attribute (weather) and wind speed attribute (wind); the weather attribute has seven attribute values: sunny, cloudy, shower, light rain, moderate rain, heavy rain , snowing; the wind speed attribute has four attribute values: the wind force is less than 3, the wind force is 3-4, the wind force is 4-5, and the wind force is greater than 5; analyze the impact of the above attributes on the number of bikes (bikeNum), and will borrow a bike Quantity is also regarded as an attribute; these attributes can be divided into two categories: numerical attributes (avgTemp, bikeNum) and category attributes (is_holiday, weather, wind); the attribute values of category attributes are discrete and only include certain specific values. The attribute values of numeric attributes are continuous; Step 5.2: Design a new line- and set-based parallel coordinates component that simultaneously exhibits the characteristics of multivariate datasets with both categorical and numerical attributes. 7. A method for visual analysis of urban public bicycle system borrowing and returning mode according to claim 6, characterized in that a new parallel coordinate component based on lines and sets is designed in step 5.2, and simultaneously exhibits categories and Characteristics of multivariate datasets for numeric attributes, as follows: Step 5.2.1: Based on the attribute characteristics, draw the coordinate axes, corresponding to five attributes from left to right: avgTemp, bikeNum, weather, isHoliday, wind; draw five coordinate axes that are parallel to each other and perpendicular to the horizontal plane; the first two Axes represent numerical attributes, represented by a straight line, with corresponding coordinates on the straight line, connected by straight lines to represent the relationship between coordinate axes; the last three axes represent category attributes, represented by a rectangle, and each attribute value occupies one part of the rectangle A small section is called an axis column; the color of the axis column is used to distinguish different attribute values, and the number of axis columns is the number of values of all attribute values; an axis column is based on the proportion of the current attribute value of a certain site, is continued to be divided into sub-axis columns; two sub-axis columns are connected by a quadrilateral; For two parallel category attribute axes, in order to calculate the height of the axis column and the width of the quadrilateral; first, for a given site _statID , retrieve and generate records containing multiple factors that affect the amount of car rental, where date represents the date of a certain day: multiFac_rec＝[statID,date,avgTemp,bikeNum,weather,isHoliday,wind] {multiFac_rec} _statID indicates the multi-influencing factor record set of the statID site; retrieve the number of valid data items f _k,i,j from {multiFac_rec} _statID , where k corresponds to the site ID, and i is the number on the left axis An attribute value, j is an attribute value on the right axis; suppose the left axis represents weather, and the right axis represents isHoliday; when attribute value (i) = "sunny", attribute value (j) = "weekday", Then f _{k, i, j} represent in {multiFac_rec} _statID , the number of data items satisfying statID = k, weather = "sunny", isHoliday = "weekday"; the width of the quadrilateral connecting two sub-axis columns is determined by freq _{k, i , determined by j} , The length of each axis column on the left axis is proportional to sum_lAxis_freq, representing the frequency of occurrence of each attribute value, Similarly, the length of each column on the right axis is proportional to sum_rAxis_freq, The legend for the station name is shown above, and the legend for the axis column color is at the bottom; When two adjacent axes respectively represent the numerical attribute and the category attribute, the lines starting from the value attribute converge to the center point of the category attribute axis column; Step 5.2.2: Since it seems messy to display all the data items directly, in order to dig out the association between multiple influencing factors more clearly, provide a way to interact with components to help analysts filter data; when the mouse moves to connect two When the line is on a line with a value attribute, the line will be strengthened, and a prompt box will pop up for all the attribute values associated with the line; when the analyst selects a quadrilateral, all relevant connections will be displayed, and irrelevant connections will be displayed. hide.