Attendance Variance Predictor

Predict Disruption Windows

Input your historical attendance patterns and environmental factors. Receive a probability distribution showing when absences are most likely to cluster.

Base Attendance Rate (%) Typical range: 88–96% for urban districts

Historical Std Deviation (%) Measure from past 12 weeks of roll data

Seasonal Factor

Upcoming Event Impact

Prediction Horizon (days) Reliability drops sharply beyond day 10

Confidence Threshold (%) Higher = narrower window, lower certainty

Methodological Foundation

This predictor implements the stochastic modeling framework from Pereira et al. (2017), adapted for K-12 attendance patterns. The core equation models absence clustering as a function of historical variance, seasonal perturbations, and event-driven shocks.

ATTENDANCE(t) = μ_base − σ_hist × Φ⁻¹(P_conf) × S_season × E_event

where:
• μ_base = mean historical attendance rate
• σ_hist = standard deviation of weekly roll data
• Φ⁻¹ = inverse cumulative normal distribution
• S_season ∈ [0.85, 1.35] = seasonal adjustment factor
• E_event ∈ [0.95, 1.40] = event impact multiplier
• t = prediction horizon (1–14 days)

Worked Example:
Base Rate: 92.5% | Std Dev: 3.2% | Winter Flu (×1.35) | Field Trip Week (×1.25)

Expected Drop = 3.2 × 1.96 × 1.35 × 1.25 = 10.3%
Peak Risk: Day 4 (flu incubation period alignment)
Cluster Size: 4–6 students in 24-student cohort
Recommended Buffer: Pre-print 6 extra activity packets

Note: Reliability degrades beyond 14 days due to unmodeled variables (family emergencies, weather events, policy changes). Re-run calculation weekly with updated historical data.