Visualization of POI Category on the Dynamic Rasterized Map Tiles from Geo-Tagged Social Media (Twitter) with SZ-GAT

Date
2023-01-03
Authors
Xie, Huaze
Li, Da
Wang, Yuanyuan
Kawai , Yukiko
Contributor
Advisor
Department
Instructor
Depositor
Speaker
Researcher
Consultant
Interviewer
Annotator
Journal Title
Journal ISSN
Volume Title
Publisher
Volume
Number/Issue
Starting Page
2210
Ending Page
Alternative Title
Abstract
Spatial zooming graph attention networks (SZ-GAT) is an emerging framework to improve the quality of recommended places visualization on the map. With the advent of location sharing on social networks via mobile devices, the geographic characteristics of the user's points of interest (POIs) contain the visit history, map check-in positions, recommended places, and route plans. In the context of user-preferred POI prediction with map zooming SZ-GAT framework, we propose a visualization for raster category exploration that uses tweet user visit history to represent the POI visit popularity of the raster units. We concentrate on the performance of the POI data visualized map layer zooming process and our results show that the SZ-GAT framework has a better performance of raster category regression with the baselines. Raster category prediction will be used for urban area division, dynamic category feature extraction with user visit history, and government policy-making based on user behaviors of map tiles. This study promotes the progress of deep learning and data mining in the field of human geographic information.
Description
Keywords
Data Analytics, Data Mining, and Machine Learning for Social Media, graph neural network, map rasterization, map zooming prediction, user preference
Citation
Extent
10
Format
Geographic Location
Time Period
Related To
Proceedings of the 56th Hawaii International Conference on System Sciences
Table of Contents
Rights
Attribution-NonCommercial-NoDerivatives 4.0 International
Rights Holder
Local Contexts
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.