Research on Online Digital Cultures-Community Extraction and Analysis by Markov and k-Means Clustering

TY - CHAP

T1 - Research on Online Digital Cultures-Community Extraction and Analysis by Markov and k-Means Clustering

AU - Pybus, Jennifer

AU - Cote, Mark Edward

AU - Blanke, Tobias

AU - Giles Greenway

PY - 2017/9/18

Y1 - 2017/9/18

N2 - We investigate approaches to personal data analytics that involves the participation of all actors in our shared digital culture. We analyse their communities by identifying and clustering social relations using mobile and social media data. The work is part of our effort to develop tools to create a “social data commons”, an open research environment that will share innovative tools and data sets to researchers interested in accessing the data that surrounds the production and circulation of digital culture and their actors. This experiment focuses on the groups of clustered relations that are formed within a user’s social data traces. Community extraction is a popular part of the analysis of social data. We have applied the technique of Markov Clustering to the Twitter networks of social actors. Qualitatively, we demonstrate that it is more effective than the Louvain method for finding social groups known to the subjects, while still being very simple to implement. We also demonstrate that traces of cell towers captured using our “MobileMiner” mobile application are sufficient to capture significant details about their social relations by the simple application of k-means.

AB - We investigate approaches to personal data analytics that involves the participation of all actors in our shared digital culture. We analyse their communities by identifying and clustering social relations using mobile and social media data. The work is part of our effort to develop tools to create a “social data commons”, an open research environment that will share innovative tools and data sets to researchers interested in accessing the data that surrounds the production and circulation of digital culture and their actors. This experiment focuses on the groups of clustered relations that are formed within a user’s social data traces. Community extraction is a popular part of the analysis of social data. We have applied the technique of Markov Clustering to the Twitter networks of social actors. Qualitatively, we demonstrate that it is more effective than the Louvain method for finding social groups known to the subjects, while still being very simple to implement. We also demonstrate that traces of cell towers captured using our “MobileMiner” mobile application are sufficient to capture significant details about their social relations by the simple application of k-means.

UR - https://www.scopus.com/pages/publications/85038210279

U2 - 10.1007/978-3-319-71970-2_10

DO - 10.1007/978-3-319-71970-2_10

M3 - Conference paper

AN - SCOPUS:85038210279

SN - 978-3-319-71969-6

VL - 10708

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 110

EP - 121

BT - Personal Analytics and Privacy

PB - Springer Verlag

T2 - 1st International Workshop on Personal Analytics and Privacy, PAP 2017, Held in Conjunction with the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2017

Y2 - 18 September 2017 through 18 September 2017

ER -