Author:
Samuel Lee Toepke
Affiliation:
Private Engineering Firm, United States
Keyword(s):
Population Estimation, Social Media, Enterprise Architecture, Volunteered Geographic Data, Twitter, Amazon Web Services, Normalization.
Abstract:
When using social media data for population estimations, data density is of primary concern. A high density of quality, crowd-sourced data in a specified geographic area leads to a more precise estimation. Nonetheless, data acquisition/storage has to be balanced against the provisioned cost/size constraints of the technical implementation and the ability to receive data in that area. This investigation compares hourly population estimations based on Tweet quantity, for several major west coast cities in the United States of America. An estimation baseline is established, and data is artificially removed from the estimation to explore the importance of data density. Experimental data is obtained and stored using an enterprise cloud solution, density observations/results are discussed, and follow-on work is described.