Scope: Semester 1, Quarter 1, AY 2025–26. Middle School Division (Grades 6–8), Periods 2–5. This document records all data sources, derivation logic, and analyst judgment calls behind the SMEDS Space Utilization pilot report. It is the authoritative reference for tracing any metric value back to its source.

1. Data Sources & Files

Source File(s) Provided by Date received
MS master schedule (2-week grid) 00_raw_data/smeds_ms_schedule_2025-26_raw.xlsx Kristen (SMEDS) Jan 2026
Section enrollment 00_raw_data/smeds_ms_course_section_ENROLLMENTS.xls (SIS export) Matt Murphy (SMEDS IT) Jan 21, 2026
Room floor areas (sqft) Email Kristen Jan 16, 2026
Grade 5 room sharing Email Kristen Mar 2, 2026
Specialist teacher schedules 00_raw_data/teachers_schedules/Music Craig.pdf, Art Kenney.pdf, Religious Studies McLaughlin.pdf, Hill_5th_grade_Q1.pdf Kristen (SMEDS) Mar 2026

Source PDFs and original school-provided files are available in the shared Google Drive folder (00_raw_data/).

2. Join Coverage

The master schedule grid was joined to SIS enrollment records by section identifier.

Join result Count
Sections in grid 50
Sections matched to enrollment 50
Match rate 100%

All sections matched. No enrollment records were imputed or dropped. The SIS export was provided January 21, 2026; enrollment may have changed between that date and the Q1 end (October 24, 2025 → the Q1 period preceded the export by several months, so the export reflects end-of-semester state rather than a mid-quarter snapshot).

3. ATU Denominator Derivations

Baseline: 23.33 hrs/wk = 4 periods (P2–P5) × 70 min × 5 days ÷ 60. Period 1 is excluded from both numerator and denominator by design (Tuesday/Thursday only; elective and homeroom use; data not available in a format consistent with the core schedule grid).

Grade 5 sharing reductions — four rooms share scheduled time with Grade 5 (Intermediate Division). ATU denominators are reduced by the estimated hours per week unavailable to MS:

Room Baseline (hrs/wk) Reduction (hrs/wk) MS Available (hrs/wk) Confidence Basis
209 23.33 1.000 22.330 Confirmed Kristen email 2026-03-02: one 30-min session on Tues + Thurs = 1.0 hr/wk
211 23.33 2.917 20.413 Inferred Kristen email 2026-03-02: one 70-min period per A-day = 2.917 hrs/wk
218 23.33 5.833 17.497 Inferred Hill schedule (stg_specialist_hill.csv): two A-day sessions (A-P2, A-P4) = 5.833 hrs/wk
219 23.33 2.917 20.413 Inferred Kristen email 2026-03-02: one 70-min period per A-day = 2.917 hrs/wk

Falsification conditions:

  • Room 218: If Hill had only one section per A-day in Q1 (not two), the reduction = 2.917 hrs/wk and MS denominator = 20.42 hrs/wk. ATU for R218 would be lower than reported.
  • Rooms 219 and 211: The March 2 email was sent after Q1 ended. If Grade 5 sharing was a Q2-only arrangement, the Q1 denominator should be 23.33 hrs/wk for both rooms. ATU would be ~14% lower than reported.
  • Room 209: Confirmed by Kristen directly; no falsification risk beyond scheduling changes.

Assumption registry: entries dim_rooms.R218_grade5_hrs_blocked_per_wk, R219_*, R211_*, R209_* (internal analyst record).

4. Specialist Schedule Filtering Logic

Specialist teacher schedules (Art/Music/Religious Studies) cover all grade divisions. To align with the MS ATU denominator window (P2–P5), the following filters were applied when computing spec_instr_hrs_wk and implied_atu in §7 of the client report:

event_type == "Instruction"          # exclude non-teaching events
start_time >= "08:50:00"             # P2 start — excludes early-morning electives

The 08:50 cutoff excludes short elective sessions that run before P2 on some days. Including them would raise Art Room implied ATU from 69.4% to 73.7%, and Music Room from 50.2% to 54.4%. The conservative filter maintains denominator consistency with the MS grid.

Why not include all instructional hours? The specialist schedules include Lower School sessions that run outside the P2–P5 window. Including them would inflate the numerator without a corresponding denominator adjustment, producing ATU values > 100% for some rooms. The P2–P5 window is the only clean denominator available from the SMEDS data.

Source files:

Teacher Room Source
Craig (Music) Music Room 00_raw_data/teachers_schedules/Music Craig.pdf
Kenney (Art) Art Room 00_raw_data/teachers_schedules/Art Kenney.pdf
McLaughlin (Religious Studies) Religious Studies room 00_raw_data/teachers_schedules/Religious Studies McLaughlin.pdf

5. Capacity Derivation

Room capacity is derived from confirmed floor area (sqft) divided by the regulatory minimum sqft per student. Kristen confirmed floor areas by email (January 16, 2026, thread room_capacity_jan2026); she cited Florida state regulations for the per-student standard.

Standards applied:

Room type sqft/student Basis
Standard classrooms 30 Kristen email Jan 16, 2026; consistent with Florida DOE guidelines
Art Studio, Music Room 45 Kristen email Jan 21, 2026
Science Lab (Room 208 only) 40 Kristen email Jan 21, 2026

Per-room capacity table:

Room Type SqFt Capacity (used) Std (sqft/student) Capacity (check)
208 Science Lab 1176 29 40 29
209 Classroom 884 29 30 29
210 Classroom 840 28 30 28
211 Classroom 900 30 30 30
212 Classroom 900 30 30 30
213 Classroom 900 30 30 30
214 Classroom 900 30 30 30
218 Classroom 800 26 30 26
219 Classroom 736 24 30 24
220 Classroom 736 24 30 24
221 Classroom 736 24 30 24
Art Room Specialist 1170 26 45 26
Boardroom Classroom 580 19 30 19
Music Room Specialist 1365 30 45 30
Religious Studies room Specialist 756 25 45 16

Outstanding question (unanswered): Kristen was asked whether capacity represents ideal class size, fire-code maximum, or desk count. No response received. The sqft ÷ standard formula is used as the most defensible interpretation.

6. Metric Definitions (Full Formulas)

ATU — Average Time Utilization

ATU = hrs_wk_scheduled ÷ hrs_wk_available × 100

hrs_wk_scheduled  = sum of section durations in the 2-week grid ÷ 2 (to annualise to per-week)
hrs_wk_available  = baseline (23.33) − Grade5_reduction_hrs_wk (room-specific)

Benchmark: ≥75% Optimised | 60–74% Healthy | 30–59% Review | <30% Action needed

ASO — Average Seat Occupancy

ASO = enroll_wtd_avg ÷ Room_Capacity × 100

enroll_wtd_avg = Σ(Actual_Enrollment × Duration_Minutes) ÷ Σ(Duration_Minutes)

Duration-weighted to reflect seat-time consumed rather than simple mean across sessions. A 70-min session with 20 students counts twice as much as a 35-min session with 20 students.

Benchmark: 60–75% target | <50% oversized | >85% crowded

CSUI — Combined Space Utilization Index

CSUI = ATU_decimal × ASO_decimal
     = (pct_atu/100) × (pct_aso/100) × 100

Bounded above by ATU. With all rooms below 63% ATU, CSUI cannot exceed ~63% regardless of seat fill.

Benchmark: >50% Strong | 30–50% Moderate | <30% Weak

SER — Space Elasticity Ratio (school-relative quartile version)

ser_label = case_when(
  pct_capacity_trough <= Q1(pct_capacity_trough)  ~ "Underutilized (low floor)",
  pct_capacity_peak   >= Q3(pct_capacity_peak)    ~ "Under pressure (high peak)",
  pct_capacity_peak   >  Q1(pct_capacity_peak) &
    pct_capacity_trough < Q3(pct_capacity_trough) ~ "Volatile (high peak + low floor)",
  TRUE                                             ~ "Moderate"
)

pct_capacity_peak   = enroll_max / Room_Capacity × 100
pct_capacity_trough = enroll_min / Room_Capacity × 100
headroom_pct        = (Room_Capacity − enroll_max) / Room_Capacity × 100

Labels are school-relative (computed from SMEDS’s own quartile distribution, not external benchmarks). The binary ±30% pass/fail threshold is deferred pending Andrew’s review.

PDR — Peak Demand Ratio

PDR = max_blocks_any_period ÷ mean_blocks_per_period  (by day type A/B)

Counting unit: 35-min blocks — the smallest schedulable unit at SMEDS. A 70-min session counts as 2 blocks; a 35-min half-period counts as 1. Computed on the canonical A-day (Week 1 Monday) and B-day (Week 1 Tuesday) only — not summed across the 2-week grid. Grade 5 sessions in shared rooms and the Gym are excluded before counting.

A PDR well above 1.0 indicates period concentration. SMEDS A-day PDR = 1.16 (peak Period 3/4); B-day PDR = 1.12 (peak Period 2). Both near 1.0 — even distribution.

Day type Period Total blocks Mean blocks
A Period 2 11 13.75
A Period 3 16 13.75 ← peak
A Period 4 16 13.75 ← peak
A Period 5 12 13.75
B Period 2 14 12.50 ← peak
B Period 3 13 12.50
B Period 4 11 12.50
B Period 5 12 12.50 
## 
## Peak period (A-day) = Period 3 and Period 4 → PDR_A = 16 / 13.75 = **1.16**  
## 
## Peak period (B-day) = Period 2 → PDR_B = 14 / 12.50 = **1.12**

RER — Room Exclusivity Rate

RER = n_exclusive_ms_rooms ÷ n_total_ms_rooms × 100

Grouping unit = MS division. Rooms shared with Grade 5 or other divisions are included in the denominator but not the numerator.

CBI — Capacity Buffer Index (provisional)

CBI = 100 − ASO

Algebraically redundant with ASO. Retained provisionally pending Andrew’s review.

7. Metric Status Log

Metric Status Reason
ATU ✅ Computed P2–P5 only; Grade 5 sharing reductions applied per room
ASO ✅ Computed Duration-weighted enrollment (seat-time basis)
CSUI ✅ Computed ATU × ASO
CBI ✅ Computed (provisional) = 100 − ASO; confirm with Andrew whether to retain
SER ✅ Computed Based on enrollment range within 2-week Q1 snapshot; quartile-based
RER ✅ Computed Grouping unit = MS division (not department)
PDR ✅ Computed Not in original Andrew set; added as scheduling-pattern diagnostic
PSMI ⚠️ Retired Specialist rooms confirmed on-mission via teacher schedules; full-day computation not possible from MS-only grid; structural bias if computed for shared rooms
SSD ❌ Not computed Requires full school-wide room inventory — unavailable from SMEDS
IDI ❌ Excluded Algebraically redundant with ASO: both derive room capacity as sqft ÷ regulatory sqft-per-student standard (30–45 sqft/student depending on room type), so IDI (students ÷ sqft) reduces to 1/ASO up to a constant. Not an industry-standard metric with universal benchmarks.
Room clustering ⏳ Deferred Ward’s D2 hierarchical clustering on ATU, ASO, headroom%. Code complete; R208 reclassified exclusive (2026-03-25). Removed from client report pending delivery decision; add cluster membership table and dendrogram when ready.

8. Known Data Gaps

The following questions were asked of the school and received no response. They are permanently unresolved for this pilot.

Question Impact on metrics Workaround applied
What time allocation does Grade 5 have for the Boardroom? Boardroom ATU may be slightly understated if Grade 5 sessions displace available MS time Inferred MS-only based on schedule data; no LS or Grade 5 Boardroom sessions appear in the MS schedule extract; assumption logged in assumptions.json
Which specific period does Grade 5 use Room 209? Applied 1.0 hr/wk deduction based on Kristen’s confirmed session count (2 × 30 min); timing within day does not affect the deduction Confirmed via email
Is room capacity based on ideal class size, fire-code maximum, or desk count? Affects Room_Capacity and therefore ASO, CBI, SER, IDI for all rooms sqft ÷ regulatory standard (30/40/45) applied as confirmed by Kristen; basis of standard not fully resolved

SMEDS Technical Provenance  |  SpaceLogic Analytics  |  Internal use only  |  Generated 2026-03-28