Full Citation

Title: Quantifying Uncertainty in Population Weighting of Twitter Analyses for Urban Risk Assessment

Citation Type: Miscellaneous

Publication Year: 2019

ISBN:

ISSN:

DOI:

NSFID:

PMCID:

PMID:

Abstract: Twitter has increasingly been used to study various research topics such as election predictions, disease spread, etc. However, social media platforms do not saturate the entire population in a study area, especially in emerging nations, only representing more affluent subpopulations. The U.S. Army Engineer Research and Development Center, Construction Engineering Research Laboratory (ERDC-CERL), as part of a project entitled Framework for the Integration of Complex Urban Systems (FICUS), is quantifying the utility of demographic information to inform neighborhood-scale social media models. Using the example topic of infrastructure, an open-source model was constructed to collect Twitter data from the metropolitan Philippines area of Manila, geotag tweets to neighborhood grid cells based on language analysis, and produce a sentiment topic map. ERDC’s social media analysis tools incorporate quantifiable uncertainties with specific on-the-ground reporting techniques. By using the Humanitarian Crisis (HC) framework developed by PACOM (another FICUS product) as a model, a framework quantifying the likelihood of being a regular social media user was created to implement a data-driven, bottom-up framework construction nested within a knowledge-based established framework. This framework, and any other produced by the FICUS team serve as case studies for augmenting the military operational environment with quantifiable reduced uncertainties.

Url: https://apps.dtic.mil/dtic/tr/fulltext/u2/1081128.pdf

User Submitted?: No

Authors: Bastian, Elizabeth G; Myers, Natalie R; Ehlschlaeger, Charles R; Burkhalter, Jeffrey A

Publisher: US Army Corps of Engineers

Data Collections: IPUMS USA

Topics: Land Use/Urban Organization

Countries: United States