Describing temporal correlation spatially in a visual analytics environment

Abish Malik, Ross Maciejewski, Erin Hodgess, David S. Ebert

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

In generating and exploring hypotheses, analysts often want to know about the relationship between data values across time and space. Often, the analysis begins at a world level view in which the overall temporal trend of the data is analyzed and linear correlations between various factors are explored. However, such an analysis often fails to take into account the underlying spatial structure within the data. In this work, we present an interactive visual analytics system for exploring temporal linear correlations across a variety of spatial aggregations. Users can interactively select temporal regions of interest within a calendar view window. The correlation coefficient between the selected time series is automatically calculated and the resultant value is displayed to the user. Simultaneously, a linked geospatial viewing window of the data provides information on the temporal linear correlations of the selected spatial aggregation level. Linear correlation values between time series are displayed as a choropleth map using a divergent color scheme. Furthermore, the statistical significance of each linear correlation value is calculated and regions in which the correlation value falls within the 95% confidence interval are highlighted. In this manner, analysts are able to explore both the global temporal linear correlations, as well as the underlying spatial factors that may be influencing the overall trend.

Original languageEnglish (US)
Title of host publicationProceedings of the Annual Hawaii International Conference on System Sciences
DOIs
StatePublished - 2011
Externally publishedYes
Event44th Hawaii International Conference on System Sciences, HICSS-44 2010 - Koloa, Kauai, HI, United States
Duration: Jan 4 2011Jan 7 2011

Other

Other44th Hawaii International Conference on System Sciences, HICSS-44 2010
CountryUnited States
CityKoloa, Kauai, HI
Period1/4/111/7/11

Fingerprint

Time series
Agglomeration
Color

Keywords

  • Crime analysis
  • Spatial aggregation
  • Temporal correlation
  • Visual analytics

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Malik, A., Maciejewski, R., Hodgess, E., & Ebert, D. S. (2011). Describing temporal correlation spatially in a visual analytics environment. In Proceedings of the Annual Hawaii International Conference on System Sciences [5718613] https://doi.org/10.1109/HICSS.2011.144

Describing temporal correlation spatially in a visual analytics environment. / Malik, Abish; Maciejewski, Ross; Hodgess, Erin; Ebert, David S.

Proceedings of the Annual Hawaii International Conference on System Sciences. 2011. 5718613.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Malik, A, Maciejewski, R, Hodgess, E & Ebert, DS 2011, Describing temporal correlation spatially in a visual analytics environment. in Proceedings of the Annual Hawaii International Conference on System Sciences., 5718613, 44th Hawaii International Conference on System Sciences, HICSS-44 2010, Koloa, Kauai, HI, United States, 1/4/11. https://doi.org/10.1109/HICSS.2011.144
Malik A, Maciejewski R, Hodgess E, Ebert DS. Describing temporal correlation spatially in a visual analytics environment. In Proceedings of the Annual Hawaii International Conference on System Sciences. 2011. 5718613 https://doi.org/10.1109/HICSS.2011.144
Malik, Abish ; Maciejewski, Ross ; Hodgess, Erin ; Ebert, David S. / Describing temporal correlation spatially in a visual analytics environment. Proceedings of the Annual Hawaii International Conference on System Sciences. 2011.
@inproceedings{d7654280944f45b0931e2c96219dae67,
title = "Describing temporal correlation spatially in a visual analytics environment",
abstract = "In generating and exploring hypotheses, analysts often want to know about the relationship between data values across time and space. Often, the analysis begins at a world level view in which the overall temporal trend of the data is analyzed and linear correlations between various factors are explored. However, such an analysis often fails to take into account the underlying spatial structure within the data. In this work, we present an interactive visual analytics system for exploring temporal linear correlations across a variety of spatial aggregations. Users can interactively select temporal regions of interest within a calendar view window. The correlation coefficient between the selected time series is automatically calculated and the resultant value is displayed to the user. Simultaneously, a linked geospatial viewing window of the data provides information on the temporal linear correlations of the selected spatial aggregation level. Linear correlation values between time series are displayed as a choropleth map using a divergent color scheme. Furthermore, the statistical significance of each linear correlation value is calculated and regions in which the correlation value falls within the 95{\%} confidence interval are highlighted. In this manner, analysts are able to explore both the global temporal linear correlations, as well as the underlying spatial factors that may be influencing the overall trend.",
keywords = "Crime analysis, Spatial aggregation, Temporal correlation, Visual analytics",
author = "Abish Malik and Ross Maciejewski and Erin Hodgess and Ebert, {David S.}",
year = "2011",
doi = "10.1109/HICSS.2011.144",
language = "English (US)",
isbn = "9780769542829",
booktitle = "Proceedings of the Annual Hawaii International Conference on System Sciences",

}

TY - GEN

T1 - Describing temporal correlation spatially in a visual analytics environment

AU - Malik, Abish

AU - Maciejewski, Ross

AU - Hodgess, Erin

AU - Ebert, David S.

PY - 2011

Y1 - 2011

N2 - In generating and exploring hypotheses, analysts often want to know about the relationship between data values across time and space. Often, the analysis begins at a world level view in which the overall temporal trend of the data is analyzed and linear correlations between various factors are explored. However, such an analysis often fails to take into account the underlying spatial structure within the data. In this work, we present an interactive visual analytics system for exploring temporal linear correlations across a variety of spatial aggregations. Users can interactively select temporal regions of interest within a calendar view window. The correlation coefficient between the selected time series is automatically calculated and the resultant value is displayed to the user. Simultaneously, a linked geospatial viewing window of the data provides information on the temporal linear correlations of the selected spatial aggregation level. Linear correlation values between time series are displayed as a choropleth map using a divergent color scheme. Furthermore, the statistical significance of each linear correlation value is calculated and regions in which the correlation value falls within the 95% confidence interval are highlighted. In this manner, analysts are able to explore both the global temporal linear correlations, as well as the underlying spatial factors that may be influencing the overall trend.

AB - In generating and exploring hypotheses, analysts often want to know about the relationship between data values across time and space. Often, the analysis begins at a world level view in which the overall temporal trend of the data is analyzed and linear correlations between various factors are explored. However, such an analysis often fails to take into account the underlying spatial structure within the data. In this work, we present an interactive visual analytics system for exploring temporal linear correlations across a variety of spatial aggregations. Users can interactively select temporal regions of interest within a calendar view window. The correlation coefficient between the selected time series is automatically calculated and the resultant value is displayed to the user. Simultaneously, a linked geospatial viewing window of the data provides information on the temporal linear correlations of the selected spatial aggregation level. Linear correlation values between time series are displayed as a choropleth map using a divergent color scheme. Furthermore, the statistical significance of each linear correlation value is calculated and regions in which the correlation value falls within the 95% confidence interval are highlighted. In this manner, analysts are able to explore both the global temporal linear correlations, as well as the underlying spatial factors that may be influencing the overall trend.

KW - Crime analysis

KW - Spatial aggregation

KW - Temporal correlation

KW - Visual analytics

UR - http://www.scopus.com/inward/record.url?scp=79952919202&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79952919202&partnerID=8YFLogxK

U2 - 10.1109/HICSS.2011.144

DO - 10.1109/HICSS.2011.144

M3 - Conference contribution

SN - 9780769542829

BT - Proceedings of the Annual Hawaii International Conference on System Sciences

ER -