Modern tourism comprehensive statistical big data system based on fused multi-channel data

Abstract

The application discloses a modern tourism comprehensive statistical big data system based on fusion of multi-channel data, relates to the technical field of big data analysis of smart tourism, and comprises the following steps: acquiring multi-source heterogeneous data of online tourism platforms, mobile communication base stations, scenic spot gate machines and social media through a data acquisition module; a data fusion module performs cleaning and standardization on the data, eliminates structural differences and semantic ambiguity, and generates uniform standard tourism behavior data items; a grid processing module constructs a space-time grid index according to geographical coordinates and time windows, and forms a dynamic tourism behavior grid mapping diagram; a situation deduction module identifies grid units with high-density tourism behavior aggregation based on the mapping diagram, and predicts the evolution trend thereof; and a decision output module generates resource configuration suggestions accordingly. The technical effect of the application is that real-time semantic-level fusion of multi-source heterogeneous tourism data is realized, and the dynamic distribution and trend of tourism activities are finely described and deduced through space-time gridding technology.

Description

Technical Field

[0001] This invention belongs to the field of smart tourism big data analysis technology, specifically a modern comprehensive tourism statistical big data system based on the integration of multi-channel data. Background Technology

[0002] Modern comprehensive tourism statistics rely on the effective collection and analysis of multi-dimensional information on tourist behavior. Currently, relevant technical solutions are typically based on single or a few homogeneous data sources, which are insufficient when integrating heterogeneous data from multiple channels such as online consumption, mobile signaling, physical access, and online public opinion. Data from different sources differ significantly in format, semantics, and timeliness, creating technical barriers. Existing data processing methods struggle to maintain the original value and real-time nature of data while achieving structural alignment and semantic uniformity across data domains. This results in the portrayal of tourist activities being limited to specific stages, failing to create a complete behavioral profile reflecting multi-dimensional characteristics such as consumption, trajectory, verification, and evaluation.

[0003] At the level of analyzing the spatiotemporal distribution of tourism activities, existing technologies mostly employ statistical models based on fixed administrative regions or predefined scenic area boundaries as basic units. This static division method cannot adapt to the dynamic clustering effects naturally formed by tourist flows. Due to the lack of a standardized and computable minimum analytical unit for continuous spatiotemporal distribution, existing analytical frameworks struggle to precisely depict and quantify the real-time formation, expansion, contraction, and migration processes of tourist hotspots in space. When predicting the evolution trend of tourist flow, its granularity and accuracy are constrained by the rigid boundaries of the inherent analytical units. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art; To this end, the present invention proposes a modern tourism comprehensive statistical big data system based on the integration of multi-channel data, comprising: The data acquisition module collects data streams from online travel platforms (reservation transaction records), mobile communication base stations (user location signaling), scenic area gate systems (tourist verification and access), and social media platforms (tourism topic content). All of these data streams are tagged with time and space. The data fusion module performs multi-source cleaning and format standardization on the booking transaction record data stream, user location signaling data stream, tourist verification and access data stream, and tourism topic content data stream, eliminating structural differences and semantic ambiguities between different data streams, and generating standardized tourism behavior data items with unified field definitions. The grid processing module inputs the standardized tourism behavior data items into the spatiotemporal grid index construction process, and assigns each data item to a specific spatiotemporal grid unit based on geographic coordinates and time windows, forming a dynamic tourism behavior grid mapping map covering the entire region. The situational analysis module, based on the dynamic tourism behavior grid mapping map, initiates tourism hotspot mining and situational analysis tasks, identifies grid cells with high-density tourism behavior clusters, and predicts the evolution trend of tourism behavior within the grid cells in future time periods. The decision output module generates a tourism service resource allocation suggestion scheme based on the identification results and evolution trend prediction of the grid cells of high-density tourism behavior aggregation.

[0005] Furthermore, the multi-source cleaning and format standardization of the booking transaction record data stream, user location signaling data stream, tourist verification and access data stream, and tourism topic content data stream specifically includes: S1: For the aforementioned booking transaction record data stream, extract the fields of order number, product type, consumption amount, departure point, destination, and booking time; match and map the non-standardized destination name with the standardized scenic spot directory database; and convert the consumption amount into a unified currency unit. S2: For the user location signaling data stream, parse the anonymous user identifier code, base station geographic location code, and signaling trigger time in the signaling, decipher the latitude and longitude coordinates through the base station geographic location code, and remove invalid location points caused by signal drift based on the continuous signaling sequence; S3: For the tourist verification and access data stream, obtain the ticket code, verification gate number, verification passage time, and ticket type information, associate the verification gate number with the specific scenic area and entrance geographical coordinates, and mark the tourist's entry attributes according to the ticket type; S4: For the data stream containing the tourism topic content, capture the text content, posting location, posting time, and sentiment tags. Use geographic entity extraction technology to identify the names of the tourism-related locations mentioned in the text content and associate them with a standardized geographic location database. The data processed through each of steps S1 to S4 is uniformly converted into a data tuple format containing user anonymity identifiers, standard geographic location coordinates, standardized timestamps, behavior type codes, and behavior attribute values, forming the standardized tourism behavior data item.

[0006] Furthermore, the process of inputting the standardized tourism behavior data items into the spatiotemporal grid index construction process, and assigning each data item to a specific spatiotemporal grid unit based on geographic coordinates and time windows, specifically involves: Predefined spatial grid division rules divide the target geographic area into equal-sized rectangular grids according to a fixed latitude and longitude span, with each rectangular grid having a unique grid code; Predefined time window division rules, using a fixed duration as the basic time window unit, and assigning a unique time window identifier to each time window; For each standardized tourism behavior data item, determine its corresponding rectangular grid code based on its included standardized geographic location coordinates; determine its corresponding time window identifier based on its included standardized timestamps. Establish a spatiotemporal grid index table with rectangular grid encoding and time window identifier as the joint primary key; Each standardized tourism behavior data item, with its behavior type code and behavior attribute value as the core content, is stored under the corresponding composite primary key entry in the spatiotemporal grid index table to form the dynamic tourism behavior grid mapping map. Each entry in the mapping map records all tourism behavior data that occurred in a specific grid within a specific time period.

[0007] Furthermore, based on the dynamic tourism behavior grid mapping map, the task of initiating tourism hotspot mining and situation projection is carried out to identify grid cells with high-density tourism behavior clusters and predict the evolution trend of tourism behavior within the grid cells in future time periods, specifically including: Scan the spatiotemporal grid index table and aggregate and statistically analyze the data in the current time window and several adjacent past time windows according to the rectangular grid encoding. Calculate the total number of tourism behavior data items, the proportion of different types of behavior data items, and the growth rate of behavior data items over time for each rectangular grid within the statistical time range. Then, weight and combine these three indicators to obtain the real-time tourism popularity value for each rectangular grid. Set a tourism popularity threshold, and mark rectangular grids whose real-time tourism popularity values ​​exceed the threshold as grid cells with high-density tourism behavior clusters; For each grid cell with a high density cluster of labeled tourism behaviors, extract its historical time series tourism popularity value and the historical popularity dissemination data of related grid cells; Using a time series prediction model, based on the historical tourism popularity values ​​of the grid cells where high-density tourism behavior is clustered, the numerical changes in tourism popularity values ​​within a specified future period are predicted. Using a spatial propagation prediction model, based on the historical heat propagation data of the associated grid cells, the direction and intensity of the heat of the grid cells with high-density tourism activity clusters spreading to surrounding grids within a specified future time period are predicted.

[0008] Furthermore, the calculation of the total number of tourism behavior data items, the proportion of different types of behavior data items, and the growth rate of behavior data items over time for each rectangular grid within the statistical time range, and the weighted summation of these three indicators, yields the real-time tourism popularity value for each rectangular grid. Specifically: The spatiotemporal grid index table is scanned within the current time window and the adjacent N past time windows. Using the rectangular grid code as the aggregation key, the data items of all behavior type codes are accumulated to obtain the total number of tourism behavior data items for each rectangular grid within the statistical time range. The number of data items for each of the four behavior types—reservation transactions, location signaling, verification and access, and social media content—is counted within each rectangular grid. The number of data items for each type is then divided by the total number of data items in the rectangular grid to obtain the percentage of data items for each behavior type. Obtain the number of data items for each rectangular grid in the current time window and the number of data items in the previous equal-length time window. Calculate the ratio of the difference between the two counts to the number of data items in the previous time window, and use this as the growth rate of the behavioral data items over time. Preset weight coefficients are assigned to the total data item index, the percentage index of each behavior type data item, and the growth rate index, respectively, wherein the growth rate index is given a negative weight to smooth short-term fluctuations. Based on the weighting coefficients, the total amount of normalized data items, the weighted sum of the proportion of each behavior type after normalization, and the normalized growth rate of each rectangular grid are linearly weighted and combined to output a comprehensive value as the real-time tourism popularity value of the rectangular grid.

[0009] Furthermore, the step of generating a tourism service resource allocation suggestion scheme based on the identification results and evolution trend prediction of the grid cells of the high-density tourism behavior clusters specifically involves: The proposed scheme includes passenger flow management routes, temporary facility deployment points, and types of public service supplements for different grid units; Obtain the geographic boundaries, current tourism popularity values, predicted future tourism popularity values, and predicted direction and intensity of popularity diffusion for grid cells marked as high-density tourism behavior clusters; From the geographic information database, query the existing tourism service resource layer information within the grid cell and the grid cell in the predicted diffusion direction. The layer information includes road network, public transportation stations, parking capacity, location of public toilets, location of medical points, and distribution of commercial service facilities. By comparing the predicted future tourism popularity with the carrying capacity of existing tourism service resources, grid areas and resource types that generate resource gaps are identified; Based on the road network topology and the predicted heat diffusion direction, a diversion and diversion path is planned from high-heat grid cells to low-heat grid cells or reserve areas. Based on the identified resource gap type and severity, the location coordinates and facility type of temporary facility deployment points are specified within or around the grid cell. The facility types include temporary parking lots, portable toilets, and emergency medical service stations. Based on the predicted changes in the proportion of tourist behavior types, suggestions are made for adjusting the types of public services to be added. These suggestions include adding guided tours in specific languages ​​and adding payment channels for specific consumption types.

[0010] Furthermore, the system also includes a step of performing deep semantic analysis on the data stream of tourism topic content on social media platforms, which is performed after the standardized tourism behavior data items are generated: From the standardized tourism behavior data items, data items whose behavior type is coded as social media content are selected; For the text content field of the data item, perform topic model analysis to extract a set of frequently occurring tourism-related keywords; Sentiment analysis is performed on the text content fields to assign a positive, neutral, or negative sentiment polarity label to each piece of content and to calculate the sentiment intensity value. The extracted set of tourism-related keywords is associated with sentiment polarity tags and sentiment intensity values ​​to construct a tourism public opinion theme sentiment map; The tourism sentiment map is associated with the standardized geographic coordinates and standardized timestamps of the corresponding data items, and added as additional semantic layer information to the data records of the corresponding spatiotemporal grid units in the dynamic tourism behavior gridded mapping map.

[0011] Furthermore, the task of identifying and predicting tourism hotspots is further integrated with the tourism public opinion sentiment map, specifically as follows: When calculating the real-time tourism popularity value for each rectangular grid, an additional sentiment popularity correction factor is introduced. The sentiment popularity correction factor is calculated based on the average sentiment intensity value of social media content within the rectangular grid and the proportion of positive sentiment content. For grid cells with high-density clusters of tourism behavior that are marked, analyze the corresponding tourism sentiment map to identify the most discussed topics and mainstream sentiments among tourists in the area where the grid cell is located. The identified key themes and mainstream sentiments are input as influencing factors into the time series prediction model and the spatial propagation prediction model to fine-tune the predicted changes in tourism popularity and the trend of popularity diffusion. For grid units with a high density of tourism behaviors where the prevailing sentiment is significantly negative, additional suggestions for public opinion guidance and emergency intervention should be added to the tourism service resource allocation proposals generated.

[0012] Furthermore, the system also includes steps for reconstructing travel trajectories and recognizing behavioral patterns based on user location signaling data streams: From a sequence of standardized travel behavior data items belonging to the same anonymous user identifier, filter out data items whose behavior type is encoded as user location; The data items are sorted according to standardized timestamps and their standard geographic location coordinates are connected to form the preliminary spatiotemporal movement trajectory of the anonymous user. The application trajectory dwell point detection algorithm identifies coordinate points in the preliminary spatiotemporal movement trajectory where the user's dwell time exceeds a threshold, and marks them as potential points of interest or overnight dwell points. The coordinates of the potential points of interest or overnight stays are matched with the tourism point of interest database to infer the user's actual attractions visited and accommodation areas; Based on the sequence of attractions visited, the duration of stay, and the speed of movement, behavioral pattern tags are assigned to users' travel behavior. These behavioral pattern tags include in-depth tour mode, fast sightseeing mode, family leisure mode, or business travel mode. The anonymous user identifier, the inferred sequence of tourist attractions, the accommodation area, and the behavioral pattern tags are used as encrypted group behavior analysis data for subsequent analysis of group tourism patterns.

[0013] Furthermore, the system utilizes the group behavior analysis data to optimize the proposed allocation scheme for tourism service resources: Aggregate the visitor visit sequences and accommodation area data of anonymous users with the same or similar behavioral patterns to analyze the spatial activity patterns and temporal distribution characteristics of tourist groups with different behavioral patterns; The spatial activity patterns of tourist groups with different behavioral patterns are superimposed onto the dynamic tourism behavior grid mapping map to analyze the pattern composition of tourist groups within a specific grid unit. When generating resource allocation recommendations for a grid cell that represents a high-density cluster of tourist behaviors, the estimated proportion of tourist group patterns within the grid cell is taken into account. For grid units primarily catering to tourists with an in-depth tour mindset, the recommended plan focuses on increasing the allocation of resources for in-depth interpretation service points, rest facilities, and cultural experience projects. For grid units primarily serving tourists in a fast-paced sightseeing mode, the recommended solution focuses on optimizing resource allocation for fast passageways, prominent signage, and efficient transportation connections. For grid units primarily frequented by family-oriented leisure visitors, the recommended plan focuses on increasing the allocation of family-friendly facilities, parent-child activity areas, and safety and security resources.

[0014] Compared with the prior art, the beneficial effects of the present invention are: Heterogeneous data streams from online travel platforms, mobile communication base stations, scenic area gates, and social media platforms undergo multi-source cleaning and format standardization. By establishing unified field definitions and semantic mapping rules, structural differences and semantic ambiguities between different data sources are eliminated. This process transforms unstructured social content, semi-structured transaction records, streaming signaling data, and IoT access data into standardized data items with consistent time labels, spatial labels, and behavioral descriptions. It achieves real-time alignment and deep integration of data across commercial, communication, IoT, and internet domains, enabling the association and integration of information on a tourist behavior across multiple dimensions such as consumption, trajectory, entry, and emotions, constructing a complete and unified panoramic view.

[0015] Based on geographic coordinates and time windows, standardized data items are categorized into specific spatiotemporal grid units, constructing a dynamic, gridded mapping map of tourism behavior covering the entire region. The spatiotemporal grid index discretizes continuous geographic space and time series into standard computational units, replacing traditional, rigid administrative division statistical units. Based on this dynamic grid map, the system can calculate the density and intensity of tourism behavior within each grid unit in real time, accurately identifying high-density grid clusters, which represent emerging tourism hotspots. By analyzing the morphology, scale, centroid movement, and intensity changes of these grid clusters in the spatiotemporal dimensions, their future evolution trajectory and trends can be predicted. This method achieves a microscopic, dynamic, and computable characterization and prediction of tourist flow distribution and movement patterns. Attached Figure Description

[0016] Figure 1 This is a timeline diagram of the modern tourism comprehensive statistical big data system based on the fusion of multi-channel data described in this invention. Figure 2 A flowchart for constructing a spatiotemporal grid index; Figure 3 A time series forecast analysis chart of tourism popularity; Figure 4 A spatiotemporal heat map of tourism popularity; Figure 5 A time trend chart of multi-channel tourism behavior data. Detailed Implementation

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] See Figure 1The data acquisition module continuously collects real-time data streams from various channels, including booking transaction records from online travel platforms, user location signaling data from mobile communication base stations, visitor verification and access data from scenic area gate systems, and tourism-related content from social media platforms. All these data streams are tagged with both time and space. The data fusion module performs multi-source cleaning and format standardization on the received multi-source data, eliminating structural differences and semantic ambiguities between different data streams, ultimately generating standardized tourism behavior data items with unified field definitions. The gridding module receives the standardized tourism behavior data items and inputs them into the spatiotemporal grid index construction process. Based on geographic coordinates and time windows, each data item is assigned to a specific spatiotemporal grid unit, forming a dynamic tourism behavior gridded mapping map covering the entire region. The situational prediction module, based on this dynamic tourism behavior gridded mapping map, initiates tourism hotspot mining and situational prediction tasks, identifying grid units with high-density tourism behavior clusters and predicting the evolution trend of tourism behavior within these grid units in future time periods. The decision output module then generates tourism service resource allocation recommendations based on the identification results and evolution trend predictions of the grid units with high-density tourism behavior clusters.

[0019] In one embodiment of the present invention, the data fusion module performs multi-source cleaning and format standardization on the booking transaction record data stream, user location signaling data stream, tourist verification and access data stream, and tourism topic content data stream to generate standardized tourism behavior data items. In specific implementation, when processing the booking transaction record data stream, the system extracts the order number, product type, consumption amount, departure point, destination, and booking time fields from the original data stream. The destination field may contain various non-standard expressions such as "West Lake Scenic Area" or "Hangzhou West Lake Scenic Area." The system matches and maps these names to a built-in standardized scenic area directory database, uniformly mapping them to the standard name "West Lake Scenic Area." For the consumption amount field, the system performs a unified currency conversion based on the transaction currency information, converting all amounts to RMB. In some embodiments, the conversion process uses a dynamic exchange rate. For an order of USD 150, if the USD to RMB conversion factor F_x is 6.80 on that day, the converted consumption amount is recorded as RMB 1020.

[0020] In practical implementation, when processing user location signaling data streams, the system parses the anonymous user identifier, base station geolocation code, and signaling trigger time contained in each signaling message. The base station geolocation code needs to be deciphered by querying the base station geolocation code and latitude-longitude correspondence table to obtain the approximate latitude-longitude coordinates of the base station associated with the user equipment at the time of signaling trigger. Based on continuous spatiotemporal signaling sequences under the same anonymous user identifier, the system uses an algorithm based on speed and distance thresholds to eliminate invalid location points that significantly deviate from reasonable movement trajectories. For example, a signaling record that jumps from city A to city B hundreds of kilometers away in a short period of time will be identified as noise data generated by signal drift and filtered out. In some embodiments, when processing tourist verification and access data streams, the system obtains verification records from scenic area gates. These records include ticket codes, gate numbers, verification times, and ticket type information. The gate number is linked to the specific scenic area name and its precise geographical coordinates by querying the scenic area gate geographic information registry. The ticket type information is used to mark the tourist's entry attributes, such as marking "adult full-day ticket" as "full price ticket" and "senior discount ticket" as "discount ticket." In specific implementations, when processing tourism-related content data streams, the system crawls publicly available text content from social media platforms, along with its associated posting location, posting time, and sentiment tags obtained through preliminary analysis. The system applies geographic entity extraction technology to the text content fields to identify and extract tourism-related location names mentioned in the text, such as extracting "Lingyin Temple" from "praying for blessings at Lingyin Temple today." The extracted location names are then matched with a standardized geographic location database to confirm their standard coordinates.

[0021] See Figure 2 In one embodiment of the present invention, the gridding processing module inputs standardized tourism behavior data items into the spatiotemporal grid index construction process. This process assigns each standardized tourism behavior data item to a specific spatiotemporal grid unit based on geographic coordinates and time windows. In specific implementation, the system predefines spatial grid division rules, which divide the target geographic area into rectangular grids of equal size according to a fixed latitude and longitude span. For example, in the Hangzhou area, the system sets the longitude span to 0.01 degrees and the latitude span to 0.01 degrees, thereby dividing the entire area into countless rectangular grids with approximately equal side lengths. Each rectangular grid has a unique grid code generated by a specific algorithm using the latitude and longitude coordinates of its center point. The formula for generating the grid code is:

[0022] in: A unique grid code representing a rectangular grid. This represents the latitude value of the center point of the rectangular grid. The function represents the longitude value of the center point of a rectangular grid. This represents a mapping function that converts latitude and longitude coordinates into unique string or integer codes.

[0023] In practical implementation, the system also predefines time window division rules. The time window division rules use a fixed duration as the basic time window unit. For example, the system sets the basic time window unit to 1 hour, starting from 0:00:00 on the same day. The first time window is identified as "T_20240206_0000", representing the time period from 0:00 to 1:00 on February 6, 2024. The second time window is identified as "T_20240206_0100", and a unique time window identifier is assigned to each time window. In some embodiments, for each standardized tourism behavior data item, the system determines its corresponding rectangular grid code based on the standard geographic coordinates contained in the standardized tourism behavior data item. For example, if the standard geographic coordinates of a standardized tourism behavior data item are (30.2451°N, 120.1385°E), the system calculates that the coordinate point falls within a rectangular grid with a center point of (30.245°N, 120.138°E), thereby assigning it the corresponding rectangular grid code "G_30245_120138". The system then determines its corresponding time window identifier based on the standardized timestamp contained in the standardized tourism behavior data item. For example, the standardized timestamp "1707213600" (corresponding to February 6, 2024, 10:00:00UTC+8) will be assigned to the time window period from 10:00 AM to 11:00 AM represented by the time window identifier "T_20240206_1000".

[0024] It is understandable that the system establishes a spatiotemporal grid index table with a rectangular grid code and a time window identifier as a combined primary key. The spatiotemporal grid index table is a multi-dimensional data structure, and each record corresponds to a unique combination of "rectangular grid code - time window identifier". In specific implementation, each standardized tourism behavior data item is stored under the corresponding combined primary key entry in the spatiotemporal grid index table with its behavior type code and behavior attribute value as the core content. For example, a standardized tourism behavior data item from a scenic spot gate has a behavior type code of "CK" and a behavior attribute value of "full price ticket". After determining that its rectangular grid code is "G_30245_120138" and its time window identifier is "T_20240206_1000", this data item is stored in the record with the key value (G_30245_120138, T_20240206_1000) in the spatiotemporal grid index table. The list associated with this key value will be appended with an entry "CK: full price ticket". Optionally, as data continues to flow in, each entry in the spatiotemporal grid index table will accumulate aggregated information of all tourism behavior data items that occur within the corresponding time window for its corresponding rectangular grid, thereby forming a dynamic tourism behavior gridded mapping map. The dynamic tourism behavior gridded mapping map is essentially a discrete mapping representation of all-area tourism activities in the spatiotemporal dimension supported by the spatiotemporal grid index table.

[0025] In one embodiment of the present invention, the situational simulation module initiates tourism hotspot mining and situational simulation tasks based on a dynamic tourism behavior gridded mapping map, identifies grid cells with high-density tourism behavior clusters, and predicts the evolution trend of tourism behavior within the grid cells in future time periods. In specific implementation, the system scans the spatiotemporal grid index table and aggregates and statistically analyzes data within the current time window and several adjacent past time windows according to rectangular grid codes. For example, the system sets the statistical time range to the current time window and the three time windows preceding it, i.e., a total time span of four hours. The system traverses the spatiotemporal grid index table and aggregates all data entries with the same rectangular grid code but whose time window identifiers fall within the above four consecutive time periods.

[0026] In practical implementation, the system calculates the total number of tourism behavior data items, the proportion of different types of behavior data items, and the growth rate of behavior data items over time for each rectangular grid within the statistical time range. These three indicators are then weighted and combined to obtain the real-time tourism popularity value for each rectangular grid. The calculation process is as follows: The spatiotemporal grid index table is scanned within the current time window and the adjacent N past time windows. Using the rectangular grid code as the aggregation key, the data items of all behavior type codes are accumulated to obtain the total number of tourism behavior data items for each rectangular grid within the statistical time range. The number of data items for each of the four behavior types—booking transactions, location signaling, verification and access, and social media content—is counted separately within each rectangular grid. The number of each type of data item is divided by the total number of data items in the rectangular grid to obtain the proportion of each behavior type. The number of data items for each rectangular grid in the current time window, and the proportion of data items in the past N ... The number of data items in the previous equal-length window is used as the growth rate of the behavioral data item over time. The difference between these two values ​​is calculated and then divided by the ratio of the number of data items in the previous time window. Preset weighting coefficients are assigned to the total number of data items, the proportion of each behavioral type, and the growth rate. The growth rate is given a negative weight to smooth short-term fluctuations. Based on these weighting coefficients, the normalized total number of data items, the weighted sum of the normalized proportions of each behavioral type, and the normalized growth rate for each rectangular grid are linearly weighted and combined to output a comprehensive value as the real-time tourism popularity value for the rectangular grid. Real-time Tourism Popularity Value The calculation can be expressed as:

[0027] in: This indicates the real-time tourism popularity value. This represents the total number of data items after normalization. This represents the weighted sum of the proportions of each behavior type after normalization. This represents the normalized growth rate. , , These are the preset weighting coefficients assigned to the three indicators, and It is a negative value.

[0028] In some embodiments, the system sets a tourism popularity threshold and marks rectangular grids with real-time tourism popularity values ​​exceeding the threshold as grid cells with high-density tourism behavior clusters. For each marked grid cell with high-density tourism behavior clusters, the system extracts its historical time-series tourism popularity value and historical popularity propagation data of associated grid cells, which typically refer to spatially adjacent grid cells. It can be understood that the system uses a time-series prediction model to predict the numerical change of tourism popularity values ​​within a specified future period based on the historical tourism popularity values ​​of the grid cells with high-density tourism behavior clusters. In some embodiments, the system simultaneously uses a spatial propagation prediction model to predict the direction and intensity of the spread of popularity from the grid cells with high-density tourism behavior clusters to surrounding grids within a specified future period, based on the historical popularity propagation data of associated grid cells. Optionally, the time-series prediction model can be an ARIMA model, and the spatial propagation prediction model can be constructed based on cellular automata principles; the prediction results of the two models together constitute a description of the evolution trend of tourism behavior within the grid cell.

[0029] See Figure 3 This is a time-series forecasting and analysis chart of tourism popularity, generated based on the ARIMA model. It comprehensively displays the historical changes in popularity and future trend projections of core hotspot grids. This chart provides direct guidance for the allocation of tourism service resources. Before the peak of popularity arrives, temporary facilities need to be deployed in and around the target grid, and tourist flow diversion routes from that grid to lower-population areas need to be planned. The width of the confidence interval also indicates that relevant departments need to prepare contingency plans to deal with sudden fluctuations. Using the ARIMA model for time-series forecasting is a classic method for predicting dynamic indicators such as tourism popularity, suitable for capturing trend-based time-series patterns. By distinguishing between historical and future data and overlaying confidence intervals, the reliability and fluctuation range of the forecast are clearly demonstrated, meeting the readability requirements of professional analysis.

[0030] In one embodiment of the present invention, the decision output module generates a tourism service resource allocation suggestion scheme based on the identification results and evolution trend prediction of grid units with high-density tourism behavior clusters. The scheme includes passenger flow diversion paths, temporary facility deployment points, and public service supplementation types for different grid units. The system obtains the geographical boundaries, current tourism popularity value, predicted future tourism popularity value, and predicted heat diffusion direction and intensity of the grid units marked as high-density tourism behavior clusters. It queries the geographic information database for existing tourism service resource layer information within the grid units with high-density tourism behavior clusters and grid units in the predicted diffusion direction. The existing tourism service resource layer information includes road networks, public transportation stations, parking lot capacity, public restroom locations, medical point locations, and distribution of commercial service facilities. In its implementation, the system compares predicted future tourism popularity with the carrying capacity of existing tourism service resources to identify grid areas and resource types that create resource gaps. Based on the road network topology and predicted popularity diffusion direction, it plans diversion and diversion paths from high-population grid units to low-population grid units or backup areas. According to the identified resource gap type and severity, it designates the location coordinates and facility types of temporary facilities within or around grid units with high-density tourist activity. Facility types include temporary parking lots, portable toilets, and emergency medical service stations. Based on the predicted changes in the proportion of different tourist behavior types, it proposes adjustments to public service supplementation types, including adding guided tours in specific languages ​​and adding payment channels for specific consumption types. See Table 1 for an example of resource gap identification.

[0031] Table 1: Examples of Resource Gap Identification in High-Density Grid Cells Grid cell encoding Resource types Current capacity Demand forecasting Gap value Severity G_30245_120138 Parking spaces 500 800 300 high G_30245_120138 Public toilet stalls 100 180 80 middle G_30246_120139 Medical point reception capacity 50 people / hour 90 people / hour 40 people / hour high In some embodiments, the system further includes a step of performing deep semantic analysis on the tourism topic content data stream of social media platforms. This deep semantic analysis step is performed after generating standardized tourism behavior data items. From these standardized tourism behavior data items, data items whose behavior type is encoded as social media content are selected. Topic model analysis is performed on the text content field of these data items to extract a set of frequently occurring tourism-related keywords. Sentiment analysis is performed on the text content field to assign positive, neutral, or negative sentiment polarity labels to each piece of content and calculate sentiment intensity values. The extracted set of tourism-related keywords is associated with the sentiment polarity labels and sentiment intensity values ​​to construct a tourism public opinion theme sentiment graph. This tourism public opinion theme sentiment graph is then associated with the standardized geographic location coordinates and standardized timestamps of the corresponding data items as additional semantic layer information, supplementing the data records of the corresponding spatiotemporal grid units in the dynamic tourism behavior gridded mapping graph. It can be understood that the tourism hotspot mining and situation inference task is further combined with the tourism public opinion theme sentiment graph. When calculating the real-time tourism popularity value of each rectangular grid, an additional sentiment popularity correction factor is introduced. The calculation is based on the average sentiment intensity of social media content within the rectangular grid and the proportion of positive sentiment content. The formula is as follows:

[0032] in: Indicates the sentiment heat correction factor. This represents the average sentiment intensity value of all social media content within the rectangular grid. This indicates the proportion of positive sentiment content within the rectangular grid relative to the total amount of social media content. As a preset adjustment coefficient, the final corrected tourism popularity value of the rectangular grid is the original tourism popularity value multiplied by the sentiment popularity correction factor. .

[0033] In some embodiments, for grid cells with high-density clusters of tourist behavior, the system analyzes the sentiment map of tourist public opinion corresponding to the grid cells to identify the most discussed topics and mainstream sentiment tendencies within the area where the grid cells are located. The identified topics and mainstream sentiment tendencies are then input as influencing factors into time series prediction models and spatial propagation prediction models to fine-tune the predicted changes in tourism popularity and the trend of popularity diffusion. Optionally, for grid cells with high-density clusters of tourist behavior where the mainstream sentiment is significantly negative, additional suggestions for public opinion guidance and emergency intervention are added to the tourism service resource allocation proposals generated for these grid cells. These suggestions may include increasing on-site management personnel in relevant areas, establishing temporary complaint handling points, or issuing clarification information through official channels. It can be understood that by introducing the sentiment map of tourist public opinion, the system incorporates the dimensions of tourist emotions and public opinion into resource allocation decisions, ensuring that the generated proposals not only focus on the carrying capacity of physical resources but also on tourist experience and potential risks.

[0034] See Figure 4This is a spatiotemporal heat map of tourism popularity, displaying the changes in tourism popularity values ​​for different grid units during the time windows from 08:00 to 19:00 in the form of a heat matrix. The tourism popularity of all grids exhibits a daily cyclical pattern of "rising in the morning, peaking in the afternoon, and falling back in the evening," consistent with the typical temporal characteristics of tourist activity in scenic areas. The core peak occurs between 14:00 and 15:00, with grid G_30245_120138 reaching its highest popularity (approximately 90) at 14:00, making it the most prominent high-density clustered grid on the map. The peak popularity of grid G_30246_120139 lasts longer, while the overall popularity of grid G_30247_120140 is relatively low, reflecting significant differences in the intensity and temporal characteristics of tourist gathering in different areas. By analyzing the spatiotemporal distribution of peak popularity, potential congestion areas can be identified in advance, providing a basis for cross-grid tourist flow management.

[0035] In one embodiment of the present invention, the system performs the steps of travel trajectory reconstruction and behavior pattern recognition for user location signaling data streams. It filters data items whose behavior type is encoded as user location from a sequence of standardized travel behavior data items belonging to the same anonymous user identifier. The filtered data items are then sorted in ascending order according to their included standardized timestamps, and the standard geographic location coordinates contained in each data item are sequentially connected to form a preliminary spatiotemporal movement trajectory corresponding to the anonymous user identifier. In a specific implementation, a trajectory dwell point detection algorithm is applied to identify coordinate points in the preliminary spatiotemporal movement trajectory where the user's dwell time exceeds a preset threshold. The algorithm traverses each coordinate point sequence in the trajectory, calculates the time difference and spatial distance between consecutive coordinate points, and detects when a group of consecutive coordinate points clustered within a certain geographical area and the total dwell time exceeds a certain threshold. Greater than the preset time threshold At that time, the center point of the geographic area is marked as a potential point of interest or overnight stop, where This indicates the total time from when the trajectory first enters the geographical area to when it finally leaves. It is a time threshold set according to the scenario. For example, the threshold used to distinguish between a brief visit and a real tour might be set to 30 minutes.

[0036] In some embodiments, the system matches the coordinates of potential points of interest (POIs) or overnight stops with a pre-built database of tourist POIs. This database contains standard names, categories, and precise geographical boundaries of various tourist-related facilities such as attractions, hotels, and restaurants. By calculating the spatial relationship between coordinates and geographical boundaries, the system infers the attractions and accommodation areas visited by the user during their actual visit. For example, if the coordinates of an identified overnight stop fall within the geographical boundary polygon of a hotel, it is inferred that the user's accommodation area is that hotel. It can be understood that based on the sequence of attractions visited, the duration of stay at each attraction, and the speed of movement between attractions, the system labels the user's travel behavior with behavioral patterns. These behavioral pattern labels include in-depth tour patterns, fast sightseeing patterns, family leisure patterns, or business travel patterns. The labeling rules are based on preset feature combinations. For example, the characteristics of an in-depth tour pattern are an average stay of more than 2 hours at a single attraction and a small number of attractions visited per day, while the characteristics of a fast sightseeing pattern are a fast speed of movement between attractions and an average stay of less than 1 hour at a single attraction. In practice, anonymous user identifiers, inferred visitor sequences, accommodation areas, and behavioral pattern tags are associated and encrypted to form encrypted group behavior analysis data. This encrypted group behavior analysis data is stored in a dedicated analysis library for subsequent group tourism pattern analysis.

[0037] The system utilizes encrypted group behavior analysis data to optimize tourism service resource allocation recommendations. It aggregates visitor sequence and accommodation data from anonymous users with similar or identical behavioral pattern tags to analyze the spatial activity patterns and temporal distribution characteristics of tourist groups with different behavioral patterns. In some embodiments, the spatial activity patterns of tourist groups with different behavioral patterns are overlaid onto a dynamic tourism behavior grid map to analyze the pattern composition of tourist groups within a specific grid cell. For example, by querying all encrypted group behavior analysis data falling within a rectangular grid encoding range, the number of different behavioral pattern tags is counted, and the proportion of each behavioral pattern tag in the tourist group of that grid cell is calculated. Optionally, when generating resource allocation recommendations for a grid cell with high-density tourism behavior clusters, the system refers to the estimated proportion of tourist group pattern composition within the grid cell. For grid cells with high-density tourism behavior clusters dominated by tourists exhibiting in-depth tour patterns, the generated tourism service resource allocation recommendations emphasize increasing the allocation of resources for in-depth interpretation service points, rest facilities, and cultural experience projects. For grid units with high-density tourist activity, primarily driven by fast-paced sightseeing, the generated tourism service resource allocation recommendations focus on optimizing the allocation of resources such as fast-access pathways, prominent signage, and efficient transportation connections. Similarly, for grid units with high-density tourist activity, primarily driven by family leisure travelers, the recommendations emphasize increasing the allocation of family-friendly facilities, parent-child activity areas, and security resources. By introducing group behavior pattern analysis, the resource allocation recommendations are further refined from solely considering visitor density to taking into account tourist behavioral preferences, making resource allocation more targeted and effective.

[0038] See Figure 5 This is a time trend chart of multi-channel tourism behavior data, showing the changing patterns of four core data categories over time: online travel platform bookings, scenic area gate verification and access, mobile communication base station location signaling, and social media topic content. All four data categories exhibit a trend of "rising in the early stage, reaching a peak in the middle stage, and falling back in the later stage," collectively peaking around the 15th-16th hour, reflecting the concentrated surge and decline of tourism activity on the same day. The consistency of the trends across the four data categories can cross-verify the authenticity of tourism popularity, avoiding misjudgments caused by the bias of a single data point. Based on the advance growth trend of booking data, subsequent visitor flow can be predicted, allowing for advance allocation of scenic area service resources; based on the peak areas of location signaling, visitor gathering hotspots can be identified for real-time crowd control. Combined with fluctuations in social media data, it is possible to correlate with tourists' emotional feedback; for example, if a surge in negative topics occurs after the peak visitor flow, it is necessary to promptly investigate and address issues with scenic area services.

[0039] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A modern tourism comprehensive statistical big data system based on the integration of multi-channel data, characterized in that, include: The data acquisition module collects data streams from online travel platforms (reservation transaction records), mobile communication base stations (user location signaling), scenic area gate systems (tourist verification and access), and social media platforms (tourism topic content). All of these data streams are tagged with time and space. The data fusion module performs multi-source cleaning and format standardization on the booking transaction record data stream, user location signaling data stream, tourist verification and access data stream, and tourism topic content data stream, eliminating structural differences and semantic ambiguities between different data streams, and generating standardized tourism behavior data items with unified field definitions. The grid processing module inputs the standardized tourism behavior data items into the spatiotemporal grid index construction process, and assigns each data item to a specific spatiotemporal grid unit based on geographic coordinates and time windows, forming a dynamic tourism behavior grid mapping map covering the entire region. The situational analysis module, based on the dynamic tourism behavior grid mapping map, initiates tourism hotspot mining and situational analysis tasks, identifies grid cells with high-density tourism behavior clusters, and predicts the evolution trend of tourism behavior within the grid cells in future time periods. The decision output module generates a tourism service resource allocation suggestion scheme based on the identification results and evolution trend prediction of the grid cells of high-density tourism behavior aggregation.

2. The modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 1, characterized in that, The multi-source cleaning and format standardization of the booking transaction record data stream, user location signaling data stream, tourist verification and access data stream, and tourism topic content data stream specifically includes: S1: For the aforementioned booking transaction record data stream, extract the fields of order number, product type, consumption amount, departure point, destination, and booking time; match and map the non-standardized destination name with the standardized scenic spot directory database; and convert the consumption amount into a unified currency unit. S2: For the user location signaling data stream, parse the anonymous user identifier code, base station geographic location code, and signaling trigger time in the signaling, decipher the latitude and longitude coordinates through the base station geographic location code, and remove invalid location points caused by signal drift based on the continuous signaling sequence; S3: For the tourist verification and access data stream, obtain the ticket code, verification gate number, verification passage time, and ticket type information, associate the verification gate number with the specific scenic area and entrance geographical coordinates, and mark the tourist's entry attributes according to the ticket type; S4: For the data stream containing the tourism topic content, capture the text content, posting location, posting time, and sentiment tags. Use geographic entity extraction technology to identify the names of the tourism-related locations mentioned in the text content and associate them with a standardized geographic location database. The data processed through each of steps S1 to S4 is uniformly converted into a data tuple format containing user anonymity identifiers, standard geographic location coordinates, standardized timestamps, behavior type codes, and behavior attribute values, forming the standardized tourism behavior data item.

3. The modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 2, characterized in that, The process of inputting the standardized tourism behavior data items into the spatiotemporal grid index construction process, and assigning each data item to a specific spatiotemporal grid unit based on geographic coordinates and time windows, specifically involves: Predefined spatial grid division rules divide the target geographic area into equal-sized rectangular grids according to a fixed latitude and longitude span, with each rectangular grid having a unique grid code; Predefined time window division rules, using a fixed duration as the basic time window unit, and assigning a unique time window identifier to each time window; For each standardized tourism behavior data item, its corresponding rectangular grid code is determined based on the standard geographic location coordinates it contains; Determine the time window identifier to which it belongs based on the standardized timestamps it contains; Establish a spatiotemporal grid index table with rectangular grid encoding and time window identifier as the joint primary key; Each standardized tourism behavior data item, with its behavior type code and behavior attribute value as the core content, is stored under the corresponding composite primary key entry in the spatiotemporal grid index table to form the dynamic tourism behavior grid mapping map. Each entry in the mapping map records all tourism behavior data that occurred in a specific grid within a specific time period.

4. A modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 3, characterized in that, Based on the dynamic tourism behavior grid mapping map, the task of tourism hotspot mining and situation prediction is initiated to identify grid cells with high-density tourism behavior clusters and predict the evolution trend of tourism behavior within the grid cells in future time periods. Specifically, this includes: Scan the spatiotemporal grid index table and aggregate and statistically analyze the data in the current time window and several adjacent past time windows according to the rectangular grid encoding. Calculate the total number of tourism behavior data items, the proportion of different types of behavior data items, and the growth rate of behavior data items over time for each rectangular grid within the statistical time range. Then, weight and combine these three indicators to obtain the real-time tourism popularity value for each rectangular grid. Set a tourism popularity threshold, and mark rectangular grids whose real-time tourism popularity values ​​exceed the threshold as grid cells with high-density tourism behavior clusters; For each grid cell with a high density cluster of labeled tourism behaviors, extract its historical time series tourism popularity value and the historical popularity dissemination data of related grid cells; Using a time series prediction model, based on the historical tourism popularity values ​​of the grid cells where high-density tourism behavior is clustered, the numerical changes in tourism popularity values ​​within a specified future period are predicted. Using a spatial propagation prediction model, based on the historical heat propagation data of the associated grid cells, the direction and intensity of the heat of the grid cells with high-density tourism activity clusters spreading to surrounding grids within a specified future time period are predicted.

5. A modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 4, characterized in that, The calculation of the total number of tourism behavior data items, the proportion of different types of behavior data items, and the growth rate of behavior data items over time for each rectangular grid within the statistical time range, and the weighted summation of these three indicators, yields the real-time tourism popularity value for each rectangular grid. Specifically: The spatiotemporal grid index table is scanned within the current time window and the adjacent N past time windows. Using the rectangular grid code as the aggregation key, the data items of all behavior type codes are accumulated to obtain the total number of tourism behavior data items for each rectangular grid within the statistical time range. The number of data items for each of the four behavior types—reservation transactions, location signaling, verification and access, and social media content—is counted within each rectangular grid. The number of data items for each type is then divided by the total number of data items in the rectangular grid to obtain the percentage of data items for each behavior type. Obtain the number of data items for each rectangular grid in the current time window and the number of data items in the previous equal-length time window. Calculate the ratio of the difference between the two counts to the number of data items in the previous time window, and use this as the growth rate of the behavioral data items over time. Preset weight coefficients are assigned to the total data item index, the percentage index of each behavior type data item, and the growth rate index, respectively, wherein the growth rate index is given a negative weight to smooth short-term fluctuations. Based on the weighting coefficients, the total amount of normalized data items, the weighted sum of the proportion of each behavior type after normalization, and the normalized growth rate of each rectangular grid are linearly weighted and combined to output a comprehensive value as the real-time tourism popularity value of the rectangular grid.

6. A modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 4, characterized in that, The step of generating a tourism service resource allocation suggestion scheme based on the identification results and evolution trend prediction of the grid cells of high-density tourism behavior aggregation is as follows: The proposed scheme includes passenger flow management routes, temporary facility deployment points, and types of public service supplements for different grid units; Obtain the geographic boundaries, current tourism popularity values, predicted future tourism popularity values, and predicted direction and intensity of popularity diffusion for grid cells marked as high-density tourism behavior clusters; From the geographic information database, query the existing tourism service resource layer information within the grid cell and the grid cell in the predicted diffusion direction. The layer information includes road network, public transportation stations, parking capacity, location of public toilets, location of medical points, and distribution of commercial service facilities. By comparing the predicted future tourism popularity with the carrying capacity of existing tourism service resources, grid areas and resource types that generate resource gaps are identified; Based on the road network topology and the predicted heat diffusion direction, a diversion and diversion path is planned from high-heat grid cells to low-heat grid cells or reserve areas. Based on the identified resource gap type and severity, the location coordinates and facility type of temporary facility deployment points are specified within or around the grid cell. The facility types include temporary parking lots, portable toilets, and emergency medical service stations. Based on the predicted changes in the proportion of tourist behavior types, suggestions are made for adjusting the types of public services to be added. These suggestions include adding guided tours in specific languages ​​and adding payment channels for specific consumption types.

7. A modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 6, characterized in that, The system also includes a step of performing deep semantic analysis on the tourism topic content data stream of social media platforms, which is performed after the standardized tourism behavior data items are generated: From the standardized tourism behavior data items, data items whose behavior type is coded as social media content are selected; For the text content field of the data item, perform topic model analysis to extract a set of frequently occurring tourism-related keywords; Sentiment analysis is performed on the text content fields to assign a positive, neutral, or negative sentiment polarity label to each piece of content and to calculate the sentiment intensity value. The extracted set of tourism-related keywords is associated with sentiment polarity tags and sentiment intensity values ​​to construct a tourism public opinion theme sentiment map; The tourism sentiment map is associated with the standardized geographic coordinates and standardized timestamps of the corresponding data items, and added as additional semantic layer information to the data records of the corresponding spatiotemporal grid units in the dynamic tourism behavior gridded mapping map.

8. A modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 7, characterized in that, The task of identifying and analyzing tourism hotspots is further integrated with the tourism public opinion sentiment map, specifically as follows: When calculating the real-time tourism popularity value for each rectangular grid, an additional sentiment popularity correction factor is introduced. The sentiment popularity correction factor is calculated based on the average sentiment intensity value of social media content within the rectangular grid and the proportion of positive sentiment content. For grid cells with high-density clusters of tourism behavior that are marked, analyze the corresponding tourism sentiment map to identify the most discussed topics and mainstream sentiments among tourists in the area where the grid cell is located. The identified key themes and mainstream sentiments are input as influencing factors into the time series prediction model and the spatial propagation prediction model to fine-tune the predicted changes in tourism popularity and the trend of popularity diffusion. For grid units with a high density of tourism behaviors where the prevailing sentiment is significantly negative, additional suggestions for public opinion guidance and emergency intervention should be added to the tourism service resource allocation proposals generated.

9. A modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 2, characterized in that, The system also includes steps for reconstructing travel trajectories and recognizing behavioral patterns based on user location signaling data streams: From a sequence of standardized travel behavior data items belonging to the same anonymous user identifier, filter out data items whose behavior type is encoded as user location; The data items are sorted according to standardized timestamps and their standard geographic location coordinates are connected to form the preliminary spatiotemporal movement trajectory of the anonymous user. The application trajectory dwell point detection algorithm identifies coordinate points in the preliminary spatiotemporal movement trajectory where the user's dwell time exceeds a threshold, and marks them as potential points of interest or overnight dwell points. The coordinates of the potential points of interest or overnight stays are matched with the tourism point of interest database to infer the user's actual attractions visited and accommodation areas; Based on the sequence of attractions visited, the duration of stay, and the speed of movement, behavioral pattern tags are assigned to users' travel behavior. These behavioral pattern tags include in-depth tour mode, fast sightseeing mode, family leisure mode, or business travel mode. The anonymous user identifier, the inferred sequence of tourist attractions, the accommodation area, and the behavioral pattern tags are used as encrypted group behavior analysis data for subsequent analysis of group tourism patterns.

10. A modern tourism comprehensive statistical big data system based on the fusion of multi-channel data as described in claim 9, characterized in that, The system utilizes the group behavior analysis data to optimize tourism service resource allocation recommendations. Aggregate the visitor visit sequences and accommodation area data of anonymous users with the same or similar behavioral patterns to analyze the spatial activity patterns and temporal distribution characteristics of tourist groups with different behavioral patterns; The spatial activity patterns of tourist groups with different behavioral patterns are superimposed onto the dynamic tourism behavior grid mapping map to analyze the pattern composition of tourist groups within a specific grid unit. When generating resource allocation recommendations for a grid cell that represents a high-density cluster of tourist behaviors, the estimated proportion of tourist group patterns within the grid cell is taken into account. For grid units primarily catering to tourists with an in-depth tour mindset, the recommended plan focuses on increasing the allocation of resources for in-depth interpretation service points, rest facilities, and cultural experience projects. For grid units primarily serving tourists in a fast-paced sightseeing mode, the recommended solution focuses on optimizing resource allocation for fast passageways, prominent signage, and efficient transportation connections. For grid units primarily frequented by family-oriented leisure visitors, the recommended plan focuses on increasing the allocation of family-friendly facilities, parent-child activity areas, and safety and security resources.

Images