STRATEGY FOR EXTRACTION OF FOURSQUARE’S SOCIAL MEDIA GEOGRAPHIC INFORMATION THROUGH DATA MINING

Abstract This aim of this paper is the acquisition of geographic data from the Foursquare application, using data mining to perform exploratory and spatial analyses of the distribution of tourist attraction and their density distribution in Rio de Janeiro city. Thus, in accordance with the Extraction, Transformation, and Load methodology, three research algorithms were developed using a tree hierarchical structure to collect information for the categories of Museums, Monuments and Landmarks, Historic Sites, Scenic Lookouts, and Trails, in the foursquare database. Quantitative analysis was performed of check-ins per neighborhood of Rio de Janeiro city, and kernel density (hot spot) maps were generated The results presented in this paper show the need for the data filtering process - less than 50% of the mined data were used, and a large part of the density of the Museums, Historic Sites, and Monuments and Landmarks categories is in the center of the city; while the Scenic Lookouts and Trails categories predominate in the south zone. This kind of analysis was shown to be a tool to support the city's tourist management in relation to the spatial localization of these categories, the tourists’ evaluations of the places, and the frequency of the target public.摘要 本研究旨在从Foursquare应用（Foursquare）获取地理数据，通过数据挖掘方法对里约热内卢市旅游景点的分布及其密度特征开展探索性分析与空间分析。据此，本研究遵循抽取-转换-加载（Extraction, Transformation, and Load，以下简称ETL）方法论，基于树状层级结构开发了三类研究算法，用于采集Foursquare数据库中博物馆、纪念地与地标、历史遗迹、观景台及步道类别的相关信息。本研究对里约热内卢各街区的签到数据开展了定量分析，并生成了核密度（热点）分布图。本研究结果表明，数据过滤流程存在必要性——仅不到50%的挖掘数据得到有效利用；博物馆、历史遗迹、纪念地与地标类别的密度主要集聚于城市中心区域，而观景台与步道类别的密度则以南片区为主。此类分析可作为支撑城市旅游管理的有效工具，能够助力相关主体掌握各类景点的空间分布、游客对景点的评价情况以及目标客群的到访频次。