School Enrollment Projections Case Study: NYC Call for Innovations METHODOLOGY EXPLANATION THOMAS MOORE APRIL 21, 2017
School Enrollment Projections Case Study:NYC Call for InnovationsMETHODOLOGY EXPLANATIONTHOMAS MOOREAPRIL 21, 2017
INTRO
NET MIGRATION
http://www.netmigration.wisc.edu/
New York, New York
Peak Age: 25-29
Net Migration
Peak Age: 25-29
Net Migration
Cook County, Illinois (Chicago)
Gwinnett County, GA (Suburb of Atlanta)
Peak Age: 30-34
Alachua County, FL (College Town)
Peak Age: 20-24
Leon County, FL (College Town)
Peak Age: 20-24
St Lucie County, FL (Typical Florida)
Pasco County
Volusia County
Lee County
Palm Beach County
Peak Age: 20-24
NET MIGRATION
INCOMING TAX FILERS TO ORANGE COUNTY, FL 2014 TO 2015
STEP 2
STEP 2: Applicants were provided a tabular Datafile:• School District 20• K-5 students, by Grade, by census tract• 10- year history (2001-02 to 2010-11)
Task:• Provide projections by each ES grade, by
census tract• 6-Year projections: Each year starting with
2011-12 through 2016-17• Similar data format
DISTRICT 20 Characteristics:
STEP 2
DISTRICT 20 HISTORICAL ENROLLMENT:
STEP 2
Why is this Area’s enrollment growing?:• Natural Increase (Births>deaths) • Replacement population of families• More people are choosing public
school
STEP 2 METHODOLOGY:
1. Create regions within District 20. (Why regions?)2. Estimate Kindergarten levels entering the school
system by region and by census tract using 2010 census data.
3. Establish grade level control numbers by region as well as aggregated for district 20.
4. Project students using cohort survival by census tract and align to the control numbers by region.
STEP 2 DATA USED:
• HISTORICAL ENROLLMENT BY CENSUS TRACT (PROVIDED)
• 2010 CENSUS DATA (P14 FILE)
• NYC GIS PARCEL LAYER (PLUTO)
TASK 1: CREATE REGIONS FOR COHORT SURVIVAL CALCULATIONS
DATA SOURCE:Parcel Layer (PLUTO)
Union PLUTO residential units with census tract/student data to get #Students/Unit
STEP 2
TASK 1: CREATE REGIONS FOR COHORT SURVIVAL CALCULATIONS
DATA SOURCE:Parcel Layer (PLUTO)Students by Census TractStudent Generation Rate
STEP 2
TASK 2: ESTIMATING KINDERGARTEN LEVELS USING CENSUS DATA
STEP 2
STARTING DATA CENSUS P14
MODEL
DISTRICTWIDE & REGIONAL CONTROL NUMBERS
DISTRICT 20 HISTORICAL ENROLLMENT
REGION A• ENROLLMENT JUMP IN 2009-10• GROWTH STABILIZING AT 1400 STUDENTS PER GRADE
LEVEL
PHASE 1
PHASE 1
DISTRICT 20 HISTORICAL ENROLLMENT
REGION C• ENROLLMENT JUMP IN 2010-11• GROWTH AT THE EARLY STAGES• HIGHER PERCENTAGE OF EXISTING POPULATION
ATTENDING PUBLIC SCHOOLS
STEP 2
DISTRICT 20 PROJECTED ENROLLMENTCONTROL NUMBERS
DISTRICT 20 ENROLLMENT PROJECTIONS:
STEP 2
FINALCensus ID
STEP 3:
PART 1 (80%): 6-Year projections of Kindergarten by Census Block for all of Queens.(Provided Historical K by census block)
PART 2 (20%): 6-Year K-5 projections by ethnicity/race, and FRM by Census Tract for all of Queens(Provided historical aggregated K-5 enrollments by ethnicity/race and FRM by census tract.
STEP 3
TASK 1: CREATE REGIONS2005 2010
STEP 3: RESIDENTIAL DEVELOPMENT-NEW CONSTRUCTION
INCLUSIONARY HOUSING DESIGNATED AREAS
STEP 3: STUDENTS CHOOSING TO ATTEND
PERCENTAGE: # OF KINDERGARTEN/ # OF KINDERGARTEN AGED STUDENTS
STEP 3 METHODOLOGY: Part 1:1. Create regions.2. Estimate Kindergarten levels entering the school system
by region and by census block using 2010 census data.3. Utilized and estimated based on the trend of school
aged students choosing to attend public school.
Part 2: 1. Projected K-5 School Age Population from census data.2. Calculated existing % of Ethnicity/Race & FRM and
applied percentage to projected school age population.
(assumes ethnicities will grow in the same proportions as the school aged population)
Winner