Citation Context

...atterns between them. • Our data set is collected from Foursquare directly. Other data sets used in many current research related to Foursquare data analysis are typically gathered from Twitter, e.g. =-=[17, 12]-=-; hence, these data sets are less complete because: 1) users may not push every check-in in Foursquare to their Twitter accounts; and, 2) there are about less than 25% Foursquare users who connect the...

Citation Context

...een data collection and publication of statistics. Digital records are becoming an increasingly important source of information to study human mobility at a global scale [11], with large implications for social sciences and epidemiology. Data from a large bill-tracking website (wheresgeorge.com), and from trajectories of trackable items recorded at the website geocaching.com, have been used to infer statistical regularities about long distance human mobility. [4] More often, geo-located records have been used to evaluate spatial mobility at city or regional level. For instance, Ferrari et al. [10] analyzed urban patterns, for the city of New York, from Twitter data. Routine whereabouts have also been extracted from applications that allow people to share their locations with friends (e.g., from Google Latitude [9] and Foursquare [15]). Mobile phone data for a small sample of people have been tracked over a period of 9 months to understand frequent mobility patterns [3]. Geo-tagged pictures in Flickr [7, 8] and recommendations posted on the social network service for travelers CouchSurfing [16] have proved useful to reconstruct and improve travel itineraries for tourists. The literature...

Citation Context

...otation technique for location-based social networks, where user check-in behavior is analyzed to extract the significance of individual places and the implicit relationships among similar places. In =-=[16]-=-, Ferrari et al. study 13 million geo-tweets collected from New York Metropolitan area to understand urban crowd behavior. Their elegant probabilistic topic-model analysis uncovers patterns that highl...

Citation Context

...users [25]. A few studies have applied topic models such as LDA to non-linguistic datasets arising from location-based services such as sets of location category tags [5] and timelocation coordinates =-=[7]-=-; these studies do not use the content of messages at all. A distinguishing characteristic of data from Foursquare and similar LBSNs is that it pairs linguistic content with spatio-temporal metadata; ...

Citation Context

...ality in developing countries (Tamgno et al., 2013). Research is this area is still at a relatively early stage, but holds promise. 1 http://www.oeaw.ac.at/vid/barometer/index.html Of all the major areas of demographic research, the study of migrations has likely benefited the most from the increasingly large availability of geo-located data from Internet sources. Travel itineraries and short term-mobility have been inferred using various data sources, including geotagged pictures in Flickr (De Choudhury et al., 2010), recommendations posted on Couchsurfing (Pultar and Raubal, 2009), Twitter (Ferrari et al., 2011), Google Latitude (Ferrari and Mamei, 2011), and Foursquare (Noulas et al., 2011). Cellphone data have also been used to evaluate patterns and regularities of internal mobility, for a country, or movements within a relatively small region (Bayir et al., 2009; Gonzalez et al., 2008; Candia et al., 2008; Blumenstock, 2012; Deville et al., 2014). Trends in international flows of migrants have been estimated by tracking the locations, inferred from IP addresses, of users who repeatedly login into Yahoo! services (Zagheni and Weber, 2012; State et al., 2013). Recently, geolocated Twitter tweets hav...

Citation Context

...hered from OpenStreetMap. Other research attempts, leveraging on social platforms such as Foursquare and Twitter, focused on spotting and labeling geographic regions according to the user activities (=-=[2,5]-=-). The remainder of this paper is organized as follows. The architecture overview of our approach is presented in Sect. 2. In Sect. 3, we detail our proposed demonstration, and in Sect. 4, we conclude...

Citation Context

... has opened new opportunities to identify migrants and to follow them, in an anonymous way, over time [14]. Various types of data sources have been used to evaluate human mobility. Cellphone data have been used mainly to evaluate patterns and regularities of internal mobility for a country [3, 13, 6, 5], but also ties between countries (in terms of international calls) [4]. Travel itineraries for tourists have been inferred using geo-tagged pictures in Flickr [9] and recommendations posted on Couchsurfing [19]. Localized mobility, often within a city, has been measured using data from Twitter [12], Google Latitude [11], Foursquare [17] and public transport fare collection sensors [15, 22]. IP addresses have been used to evaluate internal mobility [18]. Analogously, recent trends in international flows of migrants have been estimated by tracking the locations, inferred from IP addresses, of users who repeatedly login into Yahoo! services [24, 23]. 3. DATA AND METHODS 3.1 Data collection and pre-processing For this project, we downloaded geolocated Twitter ‘tweets’ for about 500,000 users who have posted at least one geolocated tweet. The data set covers the period from May 2011 to April...

Citation Context

...d junction, (2) users’ cars report their real-time GPS location, and (3) Social ‘sensors’ (here we use Twitter as an example) monitor information from trusted authorities regarding traffic conditions =-=[8]-=-. As the number of cars and cityscape increases in size, it quickly becomes impossible to provide the necessary computational power and network infrastructure to feed data from all these sources to a ...

Citation Context

...din et al., 2009). Various researchers have used Flickr data to quantify ‘urban attractiveness’ or ‘hotspots’ based on digital footprints of the visitors in various locations (Girardin et al., 2008a; =-=Ferrari et al., 2011-=-; Gavricet et al., 2011). In a similar context, Popescu and Grefenstette (2009) reported a metadata filter to automatically extract geo-tagged images and estimate visit duration times to answer questi...