REDCAP: Regionalization with Constrained Clustering and Partitioning
Regionalization is to divide a large set of spatial objects into a number of spatially contiguous regions while optimizing an objective function, which is normally a homogeneity (or heterogeneity) measure of the derived regions. This software package includes a family of four hierarchical regionalization methods, which are based on three agglomerative clustering approaches, including the single linkage, average linkage (ALK), complete linkage (CLK), and the WARD's method. The Ward's, Full-Order-ALK, and the Full-Orde-CLK methods all can produce significantly better results than existing approaches. For a specific application context, it is recommended that these three methods are applied and compared.
- (2011). "Automatic Region Building for Spatial Analysis", Transactions in GIS, 15(s1), pages 29-45. [Journal Link]
- Guo, D. (2008). "Regionalization with Dynamically Constrained Agglomerative Clustering and Partitioning (REDCAP)".International Journal of Geographical Information Science. 22(7), pp. 801-823.
- (2009). "Greedy optimization for contiguity constrained hierarchical clustering", The 4th International Workshop on Spatial and Spatiotemporal Data Mining, IEEE International Conference on Data Mining, pp. 591-596.
- Guo, D., M. Gahegan, A.M. MacEachren, and B. Zhou. "Multivariate Analysis and Geovisualization with an Integrated Geographic Knowledge Discovery Approach".Cartography and Geographic Information Science,Vol. 32, No. 2, 2005, pp. 113-132.
- Guo, D., J. Chen, A. M. MacEachren, and K. Liao (2006), "A Visualization System for Spatio-Temporal and Multivariate Patterns (VIS-STAMP)", IEEE Transactions on Visualization and Computer Graphics, 12(6), pp. 1461-1474.
Java is needed to run the following software. You can verify if Java is already installed on your computer at this link: http://java.com/en/download/installed.jsp.
A new version of REDCAP with smoothing has just been released (see below).
- API methods and a manual are added, which you can use to integrate REDCAP with your program or run REDCAP in a batch mode.
| Attachment | Size |
|---|---|
| redcap.zip | 10.72 MB |
The materials distributed on this website since 2008 are based upon work partially supported by the National Science Foundation under Grant No. 0748813. Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).
Comments
distance measurement and cluster algorithm
Hello,First, thank you very much for the development of this tool.I have two questions regarding the distance measurement and the cluster algorithm in use. I hope you can help me to figure it out.First I would like to know which distance measure is used in REDCAP to create the dissimilarity matrix? I could not find any information in the documentation or publications. I'd like to use a binary distance measurement since this represents best my spatial agglomerations. Do I have to make a binary table first and the give it to REDCAP. Or is there anyway to do so in Redcap directly? Which distance measurement is default or do I have to give a dissimilarity matrix as an Input?Secondly I'm interested why you recommend ALK as the best cluster algorithm. Standard algorithm in literature is always Ward so I wonder why you don't recommend this one since it is implemented in REDCAP.Thx for your help and patience.
using REDCAP without shape-file
Hi, I'd like to test the results of REDCAP on a data I have to compare with my current partition. I have for each of the 550 geographical unit (GU) of the map:
- value of the GU (which is to be homogenized in the territories)
- weight of the GU,
- list of neighbooring units,
- coordinates of the centroid of the GU.
I don't have a shape file with ArcMap to show the map.
Is it possible to use the software with only the data I have as I am just interested in the within-territory variance that REDCAP could find?
Sincerely.
Pierre
Shape File Type
REDCAP works with Polygon shape files but not Polygon ZM. You can convert a Polygon ZM file to a Polygon shape file in ArcGIS:
Evaluation and Comparison Results with Synthetic Data
We recently conducted an evaluation of different regionalization methods with a large number of synthetic data. The results are partially available here, which is part of a paper that is currently in press with the journal Transactions in GIS. The new version of REDCAP with smoothing will be released in August, 2011.
New Release of REDCAP
The REDCAP.jar file has been updated with a slightly new version. The old version assumed that each object had no more than 10 spatial neighbors. Now each object can have any number of neighbors.
If you cannot run the JAR file for any reason, please let us know.
Tips for REDCAP
The construction of contiguity matrix and aggregation of sets of shapes to regions may take a long time if the shapes (e.g., polygons) have very detailed boundaries, which are not necessary for REDCAP since it focuses on topology (i.e., contiguity). It is highly recommended that you simplify (generalize) the shape boundaries of your data with ArcMap (version 9.3.1):
For ArcMap (version 10):
Then you will have a new shape file. Compile your data set for REDCAP based on this new shape file. This will significantly improve the performance of and your experience with REDCAP.