Fig 1.
Schematic diagram of the tourism text data granular structure at multi-spatiotemporal scales.
(a) granular structure of tourism text data; (b) popularity mining of single spatial scale data granules at a single temporal scale; (c) popularity mining of single spatial scale data granules at multi-temporal scales; (d) popularity mining of multi-spatial scale data granules at a single temporal scale; (e) popularity mining of multi-spatial scale data granules at a multi-temporal scale.
Fig 2.
Construction of multi-spatiotemporal scale datasets of tourism text based on granular computing.
(a) dataset construction in the spatial dimension; (b) data granules at different scales; (c) dataset construction in the temporal dimension. The dataset in Fig 2(a) is constructed from bottom to top in four scales, including scenic spot, tourist route, scenic area and tourist destination. Scenic spot granules at the scenic spot scale are colored yellow; and each scenic spot granule represents a scenic spot composed of three parts: a single spot, a route with multiple spots and multiple routes with multiple spots. Route granules at the tourist route scale are colored blue, and each route granule represents a route composed of five parts: single spot, one route with multiple spots, multiple routes with multiple spots, single route and multiple routes. The single spots are represented by the complete set of scenic spots in the route (i.e., union). Similarly, the same approach is used for representing one route with multiple spots and multiple routes with multiple spots. Scenic area granules at the scenic area scale are colored light blue, and each scenic area granule represents a scenic area composed of four parts: scenic spot, tourist route, scenic area and nontoponym text. Among these, scenic spots are represented by the union set of single spots, one route with multiple spots and multiple routes with multiple spots at the tourist route scale, and a tourist route is composed of the union set of single routes and multiple routes at the tourist route scale. The tourist destination granules within the scale of tourist destination are colored aqua, and each tourist destination granule represents a tourist destination, which is represented by the union of several scenic areas within the tourism destination. The dataset in Fig 2(c) is constructed using four scales—year, month, day and hour—in top-down order, and the corresponding colors are light pink, dark pink, red and yellow. The granules at each spatial scale in Fig 2(b) are divided into four temporal scales.
Table 1.
Effective microblogs in Jiuzhaigou in 2013–2017.
Fig 3.
Correlation diagram of multi-spatiotemporal scale data in Jiuzhaigou.
Fig 4.
The variation tendencies of the popularity of Jiuzhaigou at various temporal scales.
(a1) scenic areas at the monthly scale L1&2. (a2) scenic areas at the hourly scale L1&4. (a3) scenic areas for 2013 at the daily scale L1&3_2013/b/c. (a4) scenic areas for 2017 at the daily scale L1&3_2017/b/c. (b1) tourist routes at the monthly scale L1&2. (b2) tourist routes at the hourly scale L1&4. (c1) scenic spots at the monthly scale L1&2. (c2) scenic spots at the hourly scale L1&4. (d1) popularity based on the numbers of microblog posts in August 2017. (d2) popularity based on the numbers of microblog posts in Rizegou.
Fig 5.
Comprehensive popularity of scenic spots represented by text data granules at the scenic spot scale.
Table 2.
Calculated results of TDP (abbreviated pop.) at Jiuzhaigou tourist destination scales.
Table 3.
Calculated results of TDP (abbreviated pop.) at Jiuzhaigou scenic area scales.
Table 4.
Calculated results of TDP (abbreviated pop.) at Jiuzhaigou tourist route scales.
Table 5.
Calculated results of TDP (abbreviated pop.) at Jiuzhaigou scenic spot scales.