Personal activity centres and geosocial data analysis: Combining big data with small data

Understanding how people move and interact within urban settings has been greatly facilitated by the expansion of personal computing and mobile studies. Geosocial data derived from social media applications have the potential to both document how large segments of urban populations move about and us...

Full description

Saved in:
Bibliographic Details
Main Authors: Colin Robertson, Rob Feick, Martin Sykora, Ketan Shankardass, Krystelle Shaughnessy
Format: Default Conference proceeding
Published: 2017
Subjects:
Online Access:https://hdl.handle.net/2134/24909
Tags: Add Tag
No Tags, Be the first to tag this record!
id rr-article-9501464
record_format Figshare
spelling rr-article-95014642017-01-01T00:00:00Z Personal activity centres and geosocial data analysis: Combining big data with small data Colin Robertson (343946) Rob Feick (7197629) Martin Sykora (1256697) Ketan Shankardass (508693) Krystelle Shaughnessy (6940844) Other commerce, management, tourism and services not elsewhere classified Geosocial Personal activity centres Clustering Spatial analysis Business and Management not elsewhere classified Understanding how people move and interact within urban settings has been greatly facilitated by the expansion of personal computing and mobile studies. Geosocial data derived from social media applications have the potential to both document how large segments of urban populations move about and use space, as well as how they interact with their environments. In this paper we examine spatial and temporal clustering of individuals’ geosocial messages as a way to derive personal activity centres for a subset of Twitter users in the City of Toronto. We compare the two types of clustering, and for a subset of users, compare to actual self-reported activity centres. Our analysis reveals that home locations were detected within 500 m for up to 53% of users using simple spatial clustering methods based on a sample of 16 users. Work locations were detected within 500 m for 33% of users. Additionally, we find that the broader pattern of geosocial footprints indicated that 35% of users have only one activity centre, 30% have two activity centres, and 14% have three activity centres. Tweets about environment were more likely sent from locations other than work and home, and when not directed to another user. These findings indicate activity centres defined from Twitter do relate to general spatial activities, but the limited degree of spatial variability on an individual level limits the applications of geosocial footprints for more detailed analyses of movement patterns in the city. 2017-01-01T00:00:00Z Text Conference contribution 2134/24909 https://figshare.com/articles/conference_contribution/Personal_activity_centres_and_geosocial_data_analysis_Combining_big_data_with_small_data/9501464 CC BY-NC-ND 4.0
institution Loughborough University
collection Figshare
topic Other commerce, management, tourism and services not elsewhere classified
Geosocial
Personal activity centres
Clustering
Spatial analysis
Business and Management not elsewhere classified
spellingShingle Other commerce, management, tourism and services not elsewhere classified
Geosocial
Personal activity centres
Clustering
Spatial analysis
Business and Management not elsewhere classified
Colin Robertson
Rob Feick
Martin Sykora
Ketan Shankardass
Krystelle Shaughnessy
Personal activity centres and geosocial data analysis: Combining big data with small data
description Understanding how people move and interact within urban settings has been greatly facilitated by the expansion of personal computing and mobile studies. Geosocial data derived from social media applications have the potential to both document how large segments of urban populations move about and use space, as well as how they interact with their environments. In this paper we examine spatial and temporal clustering of individuals’ geosocial messages as a way to derive personal activity centres for a subset of Twitter users in the City of Toronto. We compare the two types of clustering, and for a subset of users, compare to actual self-reported activity centres. Our analysis reveals that home locations were detected within 500 m for up to 53% of users using simple spatial clustering methods based on a sample of 16 users. Work locations were detected within 500 m for 33% of users. Additionally, we find that the broader pattern of geosocial footprints indicated that 35% of users have only one activity centre, 30% have two activity centres, and 14% have three activity centres. Tweets about environment were more likely sent from locations other than work and home, and when not directed to another user. These findings indicate activity centres defined from Twitter do relate to general spatial activities, but the limited degree of spatial variability on an individual level limits the applications of geosocial footprints for more detailed analyses of movement patterns in the city.
format Default
Conference proceeding
author Colin Robertson
Rob Feick
Martin Sykora
Ketan Shankardass
Krystelle Shaughnessy
author_facet Colin Robertson
Rob Feick
Martin Sykora
Ketan Shankardass
Krystelle Shaughnessy
author_sort Colin Robertson (343946)
title Personal activity centres and geosocial data analysis: Combining big data with small data
title_short Personal activity centres and geosocial data analysis: Combining big data with small data
title_full Personal activity centres and geosocial data analysis: Combining big data with small data
title_fullStr Personal activity centres and geosocial data analysis: Combining big data with small data
title_full_unstemmed Personal activity centres and geosocial data analysis: Combining big data with small data
title_sort personal activity centres and geosocial data analysis: combining big data with small data
publishDate 2017
url https://hdl.handle.net/2134/24909
_version_ 1794749395550339072