Two Web-based Spatial Data Visualization and Mining...

1

Mapcube and MapviewTwo Web-based Spatial Data

Visualization and Mining Systems

C.T. Lu, Y. Kou, H. WangDept. of Computer Science

Virginia Tech

S. Shekhar, P. Zhang, R. LiuDept. of Computer Science

University of Minnesota

2

Introduction: Spatial Data Analysis

Discover useful information from spatial data setsFrequent and interesting spatial patterns for post processing (knowledge discovery)

Knowledge Discovery ToolsSpatial data clustering, classificationSpatial outlier detectionSpatial data visualizationSpatial association: co-location rule

Drought & fire & pine treeTime-related spatial pattern: trends, sequential patterns analysis

Land use and land cover

3

Introduction : Spatial Data Analysis

Advanced spatial analysis toolsHandle explosive growth of spatial dataAutomatically transform the processed data into useful information and knowledgeSupport of visualization and data mining processSupport interactive user control over spatial data processingEasy to utilize and disseminate results

4

Introduction

Mapcube & MapviewWeb-based spatial data analysis toolImplemented in Java + CGI + MySQLIntegrate data mining techniques with interactive graphical interfaceIdentify spatial patterns and temporal trends

5

Mapcube

A web-based spatial analytical softwareHigh performance spatial visualization system for traffic dataSupport various visualization utilities

Traffic videoHighway time mapTraffic video comparison2-D data cube map

Support outlier detection

ApplicationsDecision-making for transportation managersCommuting route selection for commutersTraffic model establishment for researchers and planners

6

Spatial-Temporal Dimensions

Available• TTD : Time of Day• TDW : Day of Week• TMY : Month of Year• S (traffic application) :

• Station, Highway, Regions

Others• Weather, Seasons, Event types, …

7

Traffic Dimensions

8

Subsets of Dimensions

TTDTDWTMYS

TTDTDWS

TTDTDW TDWS STTD

STTD TDW

9

Dimension Pair: TTD-TDW

Configuration: X-axis: Time of day; Y-axis: Day of weekf(x,y): Avg. volume over all stations for Jan 1997

Evening rush hour broader than morning rush hourRush hour starts early on Friday. Wednesday - narrower rush hour

Trends:

10

Dimension Pair: S-TTDConfiguration:

X-axis: Time of dayY-axis: Highway I-35W North Bound

f(x,y): Avg. volume over all stations for 1/15, 1997

Trends:3-Cluster• North section: Evening rush hour• Downtown area: All day rush hour• South section: Morning rush hourS-Outliers • station ranked 9th• Time: 2:35pmMissing Data

11

Dimension Pair: S-TTD

Show traffic volume of a specific station of a highway in color graphDisplay the traffic volume of a specific station during a date

12

Dimension Pair: TDW-S

Configuration:X-axis: stations on I35W South; Y-axis: day of week

f(x,y): Avg. volume over all stations for Jan-Mar 1997

Busiest segment of I-35W South is b/w Downtown MPLS & I-62Saturday has more traffic than SundayOutliers – highway branch

Trends:

13

TTDTDWS: Comparison of Traffic Video

Configuration: Traffic volume on Jan 9 (Th) and 10 (F), 1997

Trends: Evening rush hour starts earlier on FridayCongested segments: I-35W (downtown MPLS – I-62); I-94 (MPLS – St. Paul); I-494 (intersection I-35W)

14

TTDTDWTMYS(Album)

Configuration:Outer: X-axis (month of year); Y-axis (highway) Inner: X-axis (time of day); Y-axis (day of week)

Trends: Morning rush hour: I-94 East longer than I-35 W NorthEvening rush hour: I-35W North longer than I-94 EastEvening rush hour on I-94 East: Jan longer than Feb

15

MapView

A spatial analysis softwareFacilitate the observation and study of the US census data.Support the visualization of census data, as well as spatial outlier detection.Support five spatial outlier detection algorithms.

Spatial Statistics (Z-value algorithm)Moran ScatterplotScatterplotIterative Ratio AlgorithmIterative Z-value Algorithm

A Web-based GUI ApplicationDeveloped in JavaGraphical representation of US Census DataInteractive user interface

16

MapView

17

An execution example of MapViewSelect a county to display its population in 2001 and its neighbors

18

An execution example of MapViewUser input for the number of spatial outliers

19

An execution example of MapViewDisplay result in a new window and mark the detected spatial outliers

20

West Nile Virus Data SetWest Nile Virus Map (Veterinary) 2002

21

Conclusions

Huge amount of spatial data setData visualization and data mining are useful tools for discovering useful informationCube-view: a visualization software for traffic data

Provide roll-up, drill down, slice operations on spatial-temporal dimensionsObserve the summarization of spatial patterns and temporal trends

Map-view: a spatial analysis tool for census dataSupport various spatial outlier detection algorithmsDiscover abnormal spatial patterns

22

Future Directions

Spatial GranularityProvide zoom-in, zoom-out functionalitiesSupport other data warehouse operations, such as drill-down, roll-up, slice, and dice

Spatial-Temporal outlier detectionSupport mining temporal data

3D visualization: Java 3DOther Spatial Data Mining

Spatial association rule (co-location rule)

Include Other Data SetsWest Nile Virus data set, http://westnilemaps.usgs.gov/usa_avian.html

23

Future Directions

Compute and display local statisticsSpatial auto-correlationMoran’s index

Support different classification methodsEqual distanceEqual number

Other Visualization TechniquesParallel coordinatesScatterplotHistogramPie chart

24

LinksCubeview: http://europa.nvc.cs.vt.edu/~ctlu/Project/MapView/index.htmMapview: http://europa.nvc.cs.vt.edu/~ctlu/Project/MapCube/mapcube.htm

http://europa.nvc.cs.vt.edu/~ctlu/Project/MapView/Version1/index.html

25

Related Publications

Related PublicationsChang-Tien Lu, Dechang Chen, Yufeng Kou, Algorithms for Spatial Outlier Detection, IEEE International Conference on Data Mining, 2003Chang-Tien Lu, Dechang Chen, Yufeng Kou, Detecting Spatial Outliers with Multiple Attribute, IEEE International Conference on Tools with Artificial Intelligence, 2003 S. Shekhar, C. Lu, and P.Zhang. Detecting Graph-based Spatial Outlier. Intelligent Data Analysis: An International Journal, 6(5):451-468, 2002.

26

Q & A

ctlu@vt.edu