Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

2010 Hawaii Congressional Districts (use VTD not tract) #191

Merged
merged 3 commits into from
Jul 15, 2024

Conversation

christopherkenny
Copy link
Member

Redistricting requirements

Per Hawaii Revised Statutes 25-2(b)(1)-(4) and (6), as in force for the 2010 cycle, districts must:

1. not unduly favor any person or party;

2. be contiguous, except when encompassing more than one island;

3. be compact;

4. where possible, follow geographical and recognized features and coincide with tract boundaries;

6. where practicable, avoid mixing regions with different socioeconomic interests.

Algorithmic Constraints

We enforce a maximum population deviation of 0.5%. We use Census tracts in line with 25-2(b)(4). In absence of regional knowledge about features and socioeconomic interests, we use municipalities to attempt to enforce 25-2(b)(4) and (6).

Data Sources

Data for Hawaii comes from the ALARM Project's 2010 Redistricting Data Files.

Pre-processing Notes

Islands are manually connected in the adjacency graph, but this has no bearing on the simulation.

Simulation Notes

We sample 5,000 districting plans for Hawaii over 2 independent runs of the SMC algorithm. We use partial SMC to draw one district in the contiguous portion of Honolulu County and assign the remainder to district 2.

Validation

validation_20240515_2244

SMC: 5,000 sampled plans of 2 districts on 344 units
`adapt_k_thresh`=0.99 • `seq_alpha`=0.5
`pop_temper`=0

Plan diversity 80% range: 0.23 to 0.67
✖ WARNING: Low plan diversity

R-hat values for summary statistics:
   pop_overlap      total_vap       plan_dev      comp_edge    comp_polsby      pop_white 
         1.000          1.000          1.001          1.000          1.000          1.000 
      pop_aian      pop_other       pop_nhpi        pop_two      pop_black       pop_hisp 
         1.000          1.002          1.001          1.000          1.000          1.000 
     pop_asian        vap_two      vap_other       vap_nhpi      vap_asian       vap_hisp 
         1.001          1.001          1.002          1.001          1.000          1.000 
     vap_black      vap_white       vap_aian pre_20_dem_bid pre_16_dem_cli gov_18_rep_tup 
         1.000          1.001          1.000          1.000          1.000          1.000 
uss_16_dem_sch uss_16_rep_car gov_18_dem_ige pre_20_rep_tru uss_18_rep_cur pre_16_rep_tru 
         1.000          1.000          1.000          1.000          1.000          1.000 
uss_18_dem_hir         arv_20         adv_16         adv_18         arv_18         adv_20 
         1.000          1.000          1.000          1.000          1.000          1.000 
        arv_16    muni_splits            ndv            nrv        ndshare          e_dvs 
         1.000          1.001          1.000          1.000          1.000          1.000 
          egap 
         1.000 

Sampling diagnostics for SMC run 1 of 2 (2,500 samples)
         Eff. samples (%) Acc. rate Log wgt. sd  Max. unique Est. k 
Split 1     2,460 (98.4%)      9.6%        0.25 1,556 ( 98%)      5 
Resample    2,344 (93.8%)       NA%        0.25 2,242 (142%)     NA 

Sampling diagnostics for SMC run 2 of 2 (2,500 samples)
         Eff. samples (%) Acc. rate Log wgt. sd  Max. unique Est. k 
Split 1     2,462 (98.5%)      8.4%        0.25 1,597 (101%)      6 
Resample    2,351 (94.0%)       NA%        0.25 2,224 (141%)     NA 

•  Watch out for low effective samples, very low acceptance rates (less than 1%), large std.
devs. of the log weights (more than 3 or so), and low numbers of unique plans. R-hat values for
summary statistics should be between 1 and 1.05.
• Low diversity: Check for potential bottlenecks. Increase the number of samples. Examine the
diversity plot with `hist(plans_diversity(plans), breaks=24)`. Consider weakening or removing
constraints, or increasing the population tolerance. If the acceptance rate drops quickly in the
final splits, try increasing `pop_temper` by 0.01.

Checklist

  • I have followed the instructions
  • I have updated the tracker
  • All TODO lines from the template code have been removed
  • I have merged in the main branch and then recalculated summary statistics
  • I have run enforce_style() to format my code
  • The documentation copied above is up-to-date
  • There are no data files in this pull request
  • None of the file output paths (for the redist_map and redist_plans objects, and summary statistics) have been edited

@CoryMcCartan

Additional notes

  • Low # of districts and precincts + keeping main island together results in low diversity

@CoryMcCartan
Copy link
Member

The link to 2010 guidelines is broken. Let's cite the statute directly:
https://law.justia.com/codes/hawaii/2011/division1/title3/chapter25/25-2/

In 2011 the guideline was to use tracts

@CoryMcCartan CoryMcCartan added the revise documentation Requires improvements or additions to documentation label May 20, 2024
@christopherkenny
Copy link
Member Author

2010 Hawaii Congressional Districts

Redistricting requirements

Per Hawaii Revised Statutes 25-2(b)(1)-(4) and (6), as in force for the 2010 cycle, districts must:

1. not unduly favor any person or party;

2. be contiguous, except when encompassing more than one island;

3. be compact;

4. where possible, follow geographical and recognized features and coincide with tract boundaries;

6. where practicable, avoid mixing regions with different socioeconomic interests.

Algorithmic Constraints

We enforce a maximum population deviation of 0.5%. We use Census tracts in line with 25-2(b)(4). In absence of regional knowledge about features and socioeconomic interests, we use municipalities to attempt to enforce 25-2(b)(4) and (6).

Data Sources

Data for Hawaii comes from the ALARM Project's 2010 Redistricting Data Files.

Pre-processing Notes

Islands are manually connected in the adjacency graph, but this has no bearing on the simulation.

Simulation Notes

We sample 5,000 districting plans for Hawaii over 2 independent runs of the SMC algorithm. We use partial SMC to draw one district in the contiguous portion of Honolulu County and assign the remainder to district 2.

Validation

validation_20240714_1224

SMC: 5,000 sampled plans of 2 districts on 351 units
`adapt_k_thresh`=0.99 • `seq_alpha`=0.5
`pop_temper`=0           
Plan diversity 80% range: 0.093 to 0.674
✖ WARNING: Low plan diversity
R-hat values for summary statistics:
   pop_overlap      total_vap       plan_dev      comp_edge    comp_polsby      pop_white        pop_two      pop_black 
         1.000          1.000          1.000          1.000          1.000          1.000          1.000          1.000 
      pop_hisp       pop_aian      pop_other       pop_nhpi      pop_asian        vap_two      vap_other       vap_nhpi 
         1.000          1.000          1.000          1.000          1.000          1.000          1.000          1.000 
     vap_asian       vap_hisp      vap_black      vap_white       vap_aian pre_20_rep_tru pre_20_dem_bid pre_16_dem_cli 
         1.000          1.000          1.000          1.000          1.000          1.000          1.000          1.000 
gov_18_rep_tup uss_16_dem_sch uss_16_rep_car gov_18_dem_ige uss_18_rep_cur pre_16_rep_tru uss_18_dem_hir         arv_20 
         1.000          1.000          1.000          1.000          1.000          1.000          1.000          1.000 
        adv_16         adv_18         arv_18         adv_20         arv_16    muni_splits            ndv            nrv 
         1.000          1.000          1.001          1.000          1.000          1.000          1.000          1.001 
       ndshare          e_dvs           egap 
         1.000          1.000          1.001 

Sampling diagnostics for SMC run 1 of 2 (2,500 samples)
         Eff. samples (%) Acc. rate Log wgt. sd  Max. unique Est. k 
Split 1     2,451 (98.1%)     12.3%        0.29 1,591 (101%)      5 
Resample    2,322 (92.9%)       NA%        0.29 2,222 (141%)     NA 

Sampling diagnostics for SMC run 2 of 2 (2,500 samples)
         Eff. samples (%) Acc. rate Log wgt. sd  Max. unique Est. k 
Split 1     2,451 (98.0%)     10.3%        0.29 1,567 ( 99%)      6 
Resample    2,322 (92.9%)       NA%        0.29 2,194 (139%)     NA 

•  Watch out for low effective samples, very low acceptance rates (less than 1%), large std. devs. of the log weights (more than 3
or so), and low numbers of unique plans. R-hat values for summary statistics should be between 1 and 1.05.
• Low diversity: Check for potential bottlenecks. Increase the number of samples. Examine the diversity plot with
`hist(plans_diversity(plans), breaks=24)`. Consider weakening or removing constraints, or increasing the population tolerance. If
the acceptance rate drops quickly in the final splits, try increasing `pop_temper` by 0.01.

@CoryMcCartan CoryMcCartan merged commit 4ae4ed5 into main Jul 15, 2024
@CoryMcCartan CoryMcCartan deleted the HI_cd_2010 branch July 15, 2024 01:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
revise documentation Requires improvements or additions to documentation
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants