Take me out to (analyze) the ballgame : Visualization and analysis techniques for big spatial data Public Deposited

http://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/37720j172

Descriptions

Attribute NameValues
Creator
Abstract or Summary
  • For spatial data visualization, we approach two problems and provide solutions: heat map resolution selection, and heat map confidence interval presentation. Analysts often present spatial data in gridded heat maps, at some chosen resolution. However, many data types vary in density across the domain. We develop variable-resolution heat maps to visually accommodate this changing density, and an R package, varyres, to implement it. Further, heat map confidence intervals typically consist of two heat maps, one for each confidence interval bound. We develop an interactive heat map confidence interval that changes dynamically as a user moves through the interval surfaces; and an R package, mapapp, to implement it. For spatial data analysis, Bayesian hierarchical models work well for accommodat- ing complex spatial correlation structures. However, with big spatial data we face a com- putational bottleneck on the order of n3. We delineate, use, and assess three approaches to addressing the big N problem with our spatial baseball strike zone data. Hamil- tonian Monte Carlo implemented with Stan proved too slow despite computational op- timization. Predictive Process Models, a dimension reduction technique implemented with spBayes, were much faster but Markov chains failed to converge. Integrated Nested Laplace Approximation was the only successful method.
License
Resource Type
Date Available
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Non-Academic Affiliation
Keyword
Rights Statement
Publisher
Peer Reviewed
Language
Replaces
Additional Information
  • description.provenance : Made available in DSpace on 2017-09-20T22:30:05Z (GMT). No. of bitstreams: 1ComiskeyChrisW2017.pdf: 29147456 bytes, checksum: ccc975126c6889c55ecd6b27203c4cf5 (MD5)
  • description.provenance : Approved for entry into archive by Steven Van Tuyl(steve.vantuyl@oregonstate.edu) on 2017-09-20T22:30:05Z (GMT) No. of bitstreams: 1ComiskeyChrisW2017.pdf: 29147456 bytes, checksum: ccc975126c6889c55ecd6b27203c4cf5 (MD5)
  • description.provenance : Submitted by Christopher Comiskey (comiskec@oregonstate.edu) on 2017-09-08T16:54:39ZNo. of bitstreams: 1ComiskeyChrisW2017.pdf: 29147456 bytes, checksum: ccc975126c6889c55ecd6b27203c4cf5 (MD5)
  • description.provenance : Approved for entry into archive by Julie Kurtz(julie.kurtz@oregonstate.edu) on 2017-09-20T18:55:58Z (GMT) No. of bitstreams: 1ComiskeyChrisW2017.pdf: 29147456 bytes, checksum: ccc975126c6889c55ecd6b27203c4cf5 (MD5)

Relationships

In Administrative Set:
Last modified: 02/22/2018

Downloadable Content

Download PDF
Citations:

EndNote | Zotero | Mendeley

Items