Looking at the plots for "ed_explore.R", I get why the PB voters would cluster around 1 and 0 (we're looking at individuals, and some have turned out for all elections, and some have turned out for zero.
But there are also some EDs that have 0% or 100% turnout. Looking at them, it seems like these are places with really low registrations (1, 5, 10, 32 people). Are those data errors, or just exceptionally small EDs? (maybe we also don't care for the analysis were doing)
Some (but not all) of them are:
Ad 46 - Ed 037
Ad 59 - Ed 077
Ad 60 - Ed 025
Ad 65 - Ed 084
Albany - Wd 001 - Ed 005
Brookhaven - Wd 003 - Ed 217
Brookhaven - Wd 004 - Ed 278
East Fishkill - Ed 016
Guilderland - Ed 001
Hempstead - Wd 021 - Ed 110
Hempstead - Wd 021 - Ed 131
Hempstead - Wd 022 - Ed 097
Oyster Bay - Wd 015 - Ed 107
Looking at the plots for "ed_explore.R", I get why the PB voters would cluster around 1 and 0 (we're looking at individuals, and some have turned out for all elections, and some have turned out for zero.
But there are also some EDs that have 0% or 100% turnout. Looking at them, it seems like these are places with really low registrations (1, 5, 10, 32 people). Are those data errors, or just exceptionally small EDs? (maybe we also don't care for the analysis were doing)
Some (but not all) of them are: Ad 46 - Ed 037 Ad 59 - Ed 077 Ad 60 - Ed 025 Ad 65 - Ed 084
Albany - Wd 001 - Ed 005 Brookhaven - Wd 003 - Ed 217 Brookhaven - Wd 004 - Ed 278 East Fishkill - Ed 016 Guilderland - Ed 001 Hempstead - Wd 021 - Ed 110 Hempstead - Wd 021 - Ed 131 Hempstead - Wd 022 - Ed 097 Oyster Bay - Wd 015 - Ed 107