PPS boundary options: a parent’s look at what else is possible

These are best effort boundaries. NOT REAL PLANS, yet.

Independent analysis by a PPS parent, not an official Portland Public Schools product. It may contain errors; check PPS’s rightsizing scenarios and the October 6, 2026 board materials. Built with AI, directed by a PPS parent (how).

PPS has shared two scenarios, PPS A and PPS B. This site models what other options could look like, with different boundaries and the trade-offs each one makes. The goal: to encourage PPS to release more of its data and present more options. Keeping PPS A’s closures but redrawing the lines puts more kids within walking distance and keeps more schools at their size target. Under the current best example we modeled, 70.6% of elementary kids live within a 1-mile walk of their school, against 62.7% under PPS A (72.2% with no closures). On our model, all 38 neighborhood elementary schools stay at their size target every year to 2031-32, many with thin margins (see the costs below). PPS A keeps 21 of 35 at target on our model, and 18 on PPS’s own projections. (The current best counts 38 because it keeps Lent, Rigler and César Chávez as neighborhood schools with their Spanish programs.)

Elementary kids within a 1-mile walk of school

70.6%

vs 62.7% under PPS A

Neighborhood elementary schools at size target every year to 2031-32

38 of 38

vs 21 under PPS A on our model (18 on PPS’s own)

Elementary kids assigned past a closer open school

793

vs 1,970 under PPS A

The cost: kids reassigned (K-5 + 6-8)

7,036

vs 4,907 under PPS A

Also on our model (2031-32): 3,736 elementary kids live past the 1-mile bus line (4,703 under PPS A); 586 immersion students see their program move (1,567); and on PPS’s own dashboard measures, 97% of schools are sustainable (69% under PPS A), with 98% of students in them (81%).

The current best example we modeled, against PPS A. Model estimates from Census blocks, not PPS’s student records; they may not hold up on PPS’s real data. All the numbers, every plan · Method

What it costs: more moves and messier feeder paths. About 7,000 elementary and middle schoolers change schools compared with PPS Status Quo, against about 4,900 under PPS A (37% of K-8 students, against 27%: PPS’s dashboard measure, on our model). Middle-school lines are redrawn while high-school lines stay close to PPS A’s, so fewer middle schoolers go on to the same high school as most of their class: 82.1%, against 90.2% under PPS A. Three elementary schools split between two middle schools (1 under PPS A). Middle schools are held to PPS’s 500: all 13 meet it every year to 2031-32 (5 under PPS A on PPS’s figures). One high-school change, taken from the previous best: the Skyline area stays at Lincoln instead of crossing the St. Johns Bridge to Roosevelt, so Roosevelt stays in range every year to 2031-32 (falling below 1,100 from 2032-33, a year PPS doesn’t project); the gap between the largest and smallest high school grows from 362 to 424 students, and the average high-school walk from 1.93 to 1.98 miles.

What changed on 2026-10-27. The current best now keeps three more immersion programs in their own buildings: Clark Mandarin, Sitton Spanish and Ockley Green Spanish (6-8). Only programs whose buildings close, or that lose their grades, move: 4 programs and 586 students, against 820 under the previous best. The lines shift to fit: about 180 elementary kids are drawn differently (parts of Clark’s area go to Bridger and Vestal, parts of Sitton’s to Astor and Rosa Parks), and about 720 middle schoolers change schools, which ends a criss-cross of about 185 each way between Beaumont and Tubman and brings all 13 middle schools to PPS’s 500 every year (12 under the previous best). Elementary kids sent past a closer school fall from 890 to 793. The previous best (no program swaps) and the one before it (balanced middle schools) are still shown on every page.

Trade-offs

No plan wins every row: that is the trade-off.

Every number by level, with bus lines and racial mix: Options, side by side.

What the trade-offs show

Our goals

  1. Show more options. Two scenarios are two points in a wide space. We lay out more, side by side, so the trade-offs are visible.
  2. Ask the district for more options. PPS has better data than we do. Our numbers are model estimates, built from Census blocks rather than PPS’s student records, so they may not hold up on PPS’s real data. We’d like PPS to test options like these on its own data and present more than two, including lines that keep kids closer to school and schools at target every year to 2031-32.
  3. Publish the data. Our data, code and method are open. We’d like PPS to share the data behind its scenarios too: boundary files in a mapping format (not just PDFs), student counts on a privacy-safe grid, the transfer and program-follow assumptions behind each scenario, and transportation costs.

Why the modeling stops here. We could keep pushing the models: more runs, finer lines, new trade-offs to test. But past this point, more modeling adds precision the inputs can’t support. Our student counts are estimates built from 2020 Census blocks and calibrated to 2025-26 enrollment, and PPS hasn’t published the scenario-level data behind A and B: student counts on a privacy-safe grid, the transfer and program-follow assumptions under each scenario, and cost estimates. With the district’s data, the same tools could test plans far more precisely. Without it, further tuning would mostly fit our own estimates.

The aim: a community that can talk through the trade-offs with the same numbers in front of everyone.

What this site lets you do

Compare plans on a mapAny two plans, any level: see where the lines move, and tap a spot to see its school under every plan. Read a one-pager on each planWhat each plan does, its key numbers, what it does well and what it gives up. Flip between plans, or print one. See every plan side by sideWalks, school sizes, kids moved, feeders and immersion moves, for PPS’s plans and the example plans. Why 360 and 500?Where PPS’s size targets come from (the state’s Quality Education Model), and how many schools fall below them. Check school sizes, year by yearWhich schools stay at their size target to 2031-32, and any school’s projected enrollment under every plan, next to PPS’s own figures. Explore one school at a timeAny school’s actual enrollment since 2018-19, Portland State’s forecast, PPS’s own projections for each scenario, and the example plans.
How this was done and open dataHow the model works, every source, and all the data and code: How we did this · Method · Sources · Data and code

Shorter walks and more schools at target, at a cost

Why redrawn lines do better

  • PPS A’s line changes are targeted. It folds each closing school into one or two nearby schools and makes 38 targeted boundary changes (PPS scenario memo, Oct 5, 2026), but it doesn’t redraw lines around every school, so some end up too big and others too small: 20 elementary and middle schools out of range in 2031-32 on our model. PPS’s own projections show the same, such as Clark at 272 (343 with no closures) and Sunnyside, as a K-5, at 235 (the floor is 360), and George at 636 in a 623-seat building (PPS Family Guides, p. 4; capacity from PPS’s 2025-26 school profiles, second-hand).
  • PPS A adds a little to long walks; redrawing removes a lot. With today’s schools and no closures (PPS Status Quo), 1,861 elementary kids are already assigned to a farther school, past a closer open one. PPS A raises that by 109, to 1,970. The current best example we modeled cuts it to 793: 1,068 fewer than with no closures. Sending kids to the closer school also evens out sizes.
  • Size comes first. Every example plan must keep at least as many schools at target as PPS A, every year to 2031-32; within that, it finds the shortest walks.

The costs

  • Messier feeder paths. More elementary schools split between two middle schools, and fewer middle-school classes stay together into high school (numbers above).
  • More kids change schools: about 7,100 elementary and middle schoolers under the current best, against about 4,900 under PPS A (both compared with PPS Status Quo).
  • Thin size margins. Many elementary schools sit just above the 360-kid floor: in 2031-32, 17 to 22 are within 20 kids of it across the example plans (17 of 38 under the current best), against 2 of 35 under PPS A. That works if PSU’s forecast holds, as PPS assumes; if enrollment falls faster, those schools dip below 360 first.

Sources: equity_noswap_alameda_rig_hs/TRADEOFFS.md, TRADEOFFS_MATRIX.json, PPS_SCORECARD_public.md (PPS’s dashboard measures, our model) and BRIEFS.md.

The options on the table

From PPS

Example plans PPS A’s closures and programs, lines drawn differently

Options we’d like the district to present

Two scenarios are two points in a wide space. Options like these would show the trade-offs; where we modeled one, it’s linked.

Same closures, lines drawn for walking

Keep PPS A’s closures and programs; redraw elementary and middle-school lines so more kids stay close.

Tests: what the lines alone can buy.

Keep immersion programs where they are

Leave immersion programs in their buildings when the building stays open. Programs in closing buildings, or that lose their grades, still move.

The current best moves only those: Beach and James John Spanish, Rose City Park Vietnamese and César Chávez Spanish 6-8. 586 immersion students move, against 1,567 under PPS A. The previous best also moved Clark Mandarin, Sitton Spanish and Ockley Green Spanish 6-8 (820 move); the one before kept only MLK Jr Mandarin and Beaumont Spanish 6-8 in place (1,226 move).

Redraw everything together

Redraw all three levels as one system, so each school feeds cleanly into the next.

Tried, not shown It missed the size rules: the Ida B. Wells and Grant areas lack the middle schoolers for today’s buildings. Likely needs a building decision.

Rebalance PPS A

Keep PPS A’s closures and programs, and change today’s lines as little as possible while bringing schools into range.

Finding: it moves more K-5 kids than PPS A (3,594 vs 2,935), and still leaves Forest Park, Jackson and Roseway Heights short in 2031-32.

Shortest walks that keep PPS A’s count

The shortest walks that still keep at least as many schools at target as PPS A every year to 2031-32.

Tried, dropped The re-solved previous best beat it on nearly every measure, including walks.

A range, not two points

A series from fewest moves to shortest walks, so the board can see what each step costs.

Idea

A what-if, not a proposal

What if focus options were sized too?

We tested what happens to school sizes and walks if Winterhaven and da Vinci closed and their students returned to neighborhood schools. This is a sensitivity test, not a recommendation to close either school.

It rests on a second-hand count of where focus-option students live (source of the per-program numbers not stated) and assumes every returning student stays in PPS.

Built with AI. The research, data work, modeling and writing were done by AI (Claude, from Anthropic), directed and guided by Paul Frazier, a PPS parent, who made every judgment call and checked results against PPS’s documents. Best effort; it will have errors. Method · How we did this.

Thanks to the PPS Data Explorer project (pps-data) for compiled enrollment and immersion files, and the PPS Dual Language Immersion & Rightsizing project, whose snapshots supplied PPS’s 2025-26 building capacities.

Limits