This is the root page for the project management of the redistricting project. All other materials should be linked directly or indirectly through this page.



Kai, Shaowen, Yan


Wendy Cho


The planning reflects clear near-term goals and general expectation of long-term goals. near-term means in weeks; long-term means in months. For deliverables, code must be checked in SVN; documents should be uploaded to wiki and linked on wiki page

  1. Jan 10: finish documenting anomalies in near table generation at block level and initial interpretation on the anomalies
    • Deliverable: a tech report demonstrating anomalies and initial interpretation
    • Deliverable: a set of shape files, each including the polygons involved (including neighbors) in the anomaly
  2. Jan 13: comparison of polygon adjacency & bordering matrix generation methods: ArcGIS near table, ArcGIS spatial relationship, GeoDa spatial weights calculation
    • Deliverable: a tech report on accuracy/correctness, scalability, and performance for the three aforementioned methods
  3. Jan 17: finish initialization strategy coding with hole checking
    • Deliverable: hole checking algorithm and associated data structure
    • Deliverable: a revised initialization algorithm
    • Deliverable: unit test cases which at least show two scenarios, one with holes, the other without
  4. Jan 24: finish tract-level redistricting code and technical documentation, ready for code review and testing; start writing for GIScience conference paper
    • Deliverable: code in SVN
    • Deliverable: technical document on how to run the code, with sample datasets
  5. Data pre-processing
    • Handle exception 1: a unit is contained in another unit. This can be done by removing the contained unit before running the algorithm and putting it back in final solution
    • Handle exception 2: a unit has two or more polygons. What to do?
  6. March 21: finish redistricting at tract level and produce a single SVN code version with README file; document runtime environment and library dependency, as well as compiling (if needed), running, and testing instructions
  7. March 21: review exception handling to see if, for exception 2, area difference among unit polygons is large or not
  8. March 21: discussion agenda: 1) extended abstract timeline; 2) journal paper development
  9. May 04: Extended abstract for GIScience 2012 due


  • At tract level, should we do multi-scale optimization, i.e., using multi-obj pareto front at county-level to guide tract-level optimization? If yes, explain why and propose a time line; if not, explain why, too