How the Snow Day Predictor Works
snowdayprediction.ca pulls live weather data from Environment Canada every 30 minutes and runs a 9-variable scoring algorithm calibrated to the specific school board covering your city. The result is a probability score — not a guess — built from the same variables that transportation assessors evaluate at 3:00 AM on storm mornings.
A 15cm overnight snowfall forecast means something different for Toronto District School Board than it does for Simcoe County District School Board. The algorithm accounts for this. Board-specific closure thresholds, geographic footprint, rural route exposure, and historical closure frequency are all factored into every prediction.
Data Sources
Every prediction runs on live government weather data — no third-party forecasts, no aggregated commercial feeds.
Environment Canada
Hourly forecast grids, precipitation type forecasts, wind speed and direction, visibility forecasts, temperature at surface level, and all active weather alerts and warnings. Data updates every 30 minutes from Environment Canada's MSC Datamart.
Open-Meteo
High-resolution precipitation accumulation forecasts and hourly snowfall totals. Open-Meteo provides the granular overnight accumulation window data — specifically the 10pm to 6am window that school board transportation assessors use as their primary evaluation period.
School Board Calibration
Historical closure records for each school board inform the board-specific multipliers applied to the base weather score. Boards with higher historical closure frequency receive higher multipliers — the same storm produces different probability scores for different boards.
What the Tool Shows
Beyond the probability score, the tool surfaces 20+ intelligence panels built from the same underlying data.
Plain-language explanation of exactly why the score landed where it did — which factors drove it up or down.
Visual breakdown of all 9 scoring variables and their individual contribution to the final probability.
Each board has a documented closure threshold and behaviour pattern — displayed as a plain-language profile.
Four-metric breakdown: road ice risk, visibility, bus safety risk, and storm accumulation timing.
Snow day risk for each of the next 7 days — Low, Moderate, High, or Very High — with precipitation type.
How closure probability evolves hour by hour from 6 PM tonight through 7 AM tomorrow morning.
Documented historical events where this board faced similar conditions — and what they decided.
Running closure count for this board vs the historical average, with season progress and streak data.
Real-time ground reports from parents and teachers — roads icy, buses late, heavy snow, roads clear.
One-tap share to WhatsApp, iMessage, or clipboard. Parents share predictions in school group chats.
Subscribe to alerts for your city. Get notified when probability crosses 60% overnight.
Toggle between parent view (childcare focus) and teacher view (commute focus) for mode-specific guidance.
The 9 Scoring Variables
Each prediction is the weighted sum of nine variables evaluated against board-specific thresholds.
Overnight Snowfall Accumulation
High WeightTotal snowfall forecast between 10pm and 6am is the single strongest predictor of a school closure. Boards across Canada set informal thresholds — most close when overnight accumulation exceeds 15cm. The scoring engine applies board-specific thresholds based on historical closure data for each community.
Precipitation Type
High WeightFreezing rain produces closures at far lower accumulation totals than snow. A 5mm freezing rain event on rural county roads triggers closures that 15cm of snow might not. The algorithm weights freezing rain warnings from Environment Canada at 2.3x the base snowfall score.
Wind Speed and Wind Chill
Medium WeightWind speeds above 50km/h reduce visibility and create blowing snow conditions that make road assessment difficult. Wind chill below -35°C triggers closures in northern Ontario and prairie boards independently of snowfall totals. The algorithm applies a wind modifier to the base precipitation score.
Visibility Forecast
Medium WeightVisibility below 400 metres during the 6am to 9am window is a closure trigger for most boards regardless of precipitation totals. The algorithm pulls visibility forecast data from Environment Canada hourly forecast grids and applies a visibility penalty when the morning window falls below threshold.
Board-Specific Closure Multiplier
High WeightEach school board has a calibrated multiplier based on historical closure frequency. Simcoe County District School Board closes 12 to 15 times per year — the same storm that scores 40% for Toronto TDSB scores 75% for SCDSB. The multiplier is built from historical closure records and geographic footprint analysis for each board.
Rural Route Exposure
Medium WeightBoards with large rural footprints close at lower snowfall totals than urban boards. A board covering extensive county road bus routes faces road conditions that urban boards never encounter. The algorithm applies a geographic exposure factor based on the percentage of rural route coverage in each board.
Temperature Trend
Low WeightTemperature at pavement level determines whether precipitation falls as snow, freezing rain, or rain. A forecast showing -2°C at 3am rising to +1°C by 6am creates freezing rain risk regardless of precipitation type forecasts. The algorithm monitors temperature trend across the overnight assessment window.
Environment Canada Alert Status
High WeightActive Environment Canada winter weather alerts — snow squall warnings, freezing rain warnings, blizzard warnings, and winter storm watches — are weighted as high-confidence closure signals. A snow squall warning covering a school board adds a significant probability boost independent of forecast accumulation totals.
Historical Storm Pattern Match
Low WeightThe algorithm compares current forecast conditions against historical storm events that resulted in closures for each board. Pattern matching identifies combinations of variables that historically produced closures even when individual variables fell below threshold levels.
Prediction Accuracy
Accuracy varies by storm type and check time — the algorithm is most reliable at 5:00 AM.
Why marginal events are harder to predict: When overnight snowfall falls in the 5 to 12cm range, the closure decision shifts heavily toward real-time road condition assessments that no forecast model can replicate. Transportation assessors driving county roads at 3:00 AM have information that weather models cannot provide. snowdayprediction.ca reflects this uncertainty — marginal event scores of 40 to 65% accurately represent genuine ambiguity, not model failure.
When to Check the Predictor
The probability score changes as forecast data updates overnight. Two check times matter most.
Evening Check — Plan Ahead
The 9pm check gives the first reliable probability read for the following morning. Environment Canada's overnight forecast has typically stabilized by 9pm. A score above 70% at 9pm warrants backup childcare arrangements. A score below 30% at 9pm rarely results in a closure.
Morning Check — Final Read
The 5am check reflects the latest forecast data and aligns with the window when transportation assessors are actively driving routes and bus operators are making cancellation decisions. This is the most accurate pre-announcement probability available. Most Ontario boards announce by 6:00 AM.
Important Disclaimer
snowdayprediction.ca is a probability tool — not an official source. School closure decisions are made exclusively by school boards and are announced on board websites, board Twitter accounts, and local radio stations. Always verify closures through official board sources before making decisions.
snowdayprediction.ca is not affiliated with any school board, government agency, or Environment Canada. Weather data is sourced from Environment Canada MSC Datamart and Open-Meteo under open data licenses.
Ready to check tomorrow's snow day probability for your city?
Check Your City →