Using Social Network Analysis to Compare Hispanic and Black Dementia Caregiving Networks in Twitter
Video
Team Information
Team Members
Sunmoo Yoon, Associate Research Scientist, General Medicine, `Department of Medicine; and The Data Science Institute, Columbia University
Haseeb Asim, MS Candidate, Applied Analytics, School of Professional Studies, Columbia University
Nanyi Deng, MS Candidate, Applied Analytics, School of Professional Studies, Columbia University
Peter Broadwell, Research Developer, Center for Interdisciplinary Digital Research, Stanford University
Michelle Odlum, EdD, MPH, Assistant Professor, Columbia University Irving Medical Center.
Nicole J Davis, PhD, Assistant Professor, School of Nursing, Clemson University.
Carmela Alcantara, PhD, Associate Professor, School of Social Work, CU.
Mary Mittelman, Dr. PH, Professor, Department of Psychiatry, Grossman Department of Medicine, NYU
Abstract
The prevalence of Alzheimer’s disease is higher for Hispanics and Blacks than non-Hispanic Whites. Experimental evidence highlights the critical role of dyad or triad friendship relations in behavioral science. Social network analysis may provide insights to design Twitter based culturally sensitive social support intervention for Hispanic and Black family caregivers for persons with dementia. The purpose of this study is to apply social network analysis on Tweets to compare Hispanic and Black dementia caregiving networks. We randomly extracted Tweets mentioning dementia caregiving and its related terms from corpora collected daily via API from Sep 1 to Dec 31, 2019 (n= 549,380 English Tweets, n= 185,684 Spanish Tweets). We first applied a Twitter bot detection algorithm to remove bot-generated Tweets followed by applying a lexicon-based demographic inference algorithm to automatically identify Tweets likely authored by Blacks and Hispanics (n= 114,511 English, n = 1,185 Spanish). We then applied the Louvain clustering algorithm to detect groups within each Hispanic and Black caregiving network using Python and ORA. Fourteen distinct groups (11.0%, Louvain modularity: 0.80) were detected in the Hispanic caregiving network whereas 123 groups (7.0%, Louvain modularity: 0.89) were found in the Black caregiving network. Both networks contained a similar proportion of dyads and triads (Hispanics 88.2%, 88.9% Blacks) while the Black caregiving network included slightly larger proportion of isolates (Hispanics 0.8%, Blacks 4.0%). This study provides useful baseline information on the composition of existing large groups, small groups and isolates for our future recruitment strategy and design of social support intervention in regards to emotional needs for Hispanic and Black dementia caregivers.
Contact this Team
Contact: Sunmoo Yoon (use form to send email)
