Canada PR Admissions Strategic Analysis + ML Forecast
Source: IRCC Open Data · Jan 2015 – Feb 2026 · GradientBoosting model trained on 611 records
Dataset 2 files
Records 482K+ (2024)
Period 2015–2026
Model GBM · R²=0.79
Total Admissions 2024
482,960
Permanent Residents
▲ +2.5% vs 2023
Brazilian PRs 2024
5,985
Permanent Residents
▼ −17.2% vs 2023
Model R²
0.79
Variance explained
MAE = 593 admissions
Peak Quarter · Brazil
Q1
Avg 473 admissions/month
▲ 28% above Q4
Training Time
0.68s
611 records · GBM 400 trees
Test set: 474 records
Annual PR Admissions — Canada vs Brazil
Canada total (left axis) · Brazil (right axis) · 2015–2025
TIME SERIES
+208%
Brazil growth 2019→2021 pandemic rebound spike
11,415
Brazil peak admissions 2021 all-time high
−34%
Brazil decline 2021→2024 normalization post-peak
+78%
Canada growth 2015→2024 record highs consecutively
Top Immigration Categories
Cumulative 2015–2025 · all provinces
CATEGORIES
Brazil — Quarterly Seasonality
Average monthly admissions per quarter · 2015–2025
SEASONALITY
PEAK
Q1 · Jan–Mar
473
avg/month · +28% vs Q4
Q2 · Apr–Jun
439
avg/month · 2nd best
Q3 · Jul–Sep
448
avg/month · stable
Q4 · Oct–Dec
340
avg/month · weakest
Strategic Insight
Clients who begin the process in Sep–Oct land in the Q1 processing window — the strongest quarter. Advising against Q4 submissions is a measurable differentiator.
Province Distribution 2024
Top 6 provinces · share of 482,960 total
DOUGHNUT
ON+BC+AB
70% of national admissions Strategic concentration zone
Province Rankings 2024
Full breakdown with national share
TABLE
#
Province
Admissions
Share
Province Trend — Top 5 (2015–2025)
Annual admissions by major province
MULTI-LINE
Brazil — Annual PR Admissions 2015–2025
Yearly volume with growth annotation
BRAZIL
Monthly Seasonality Detail
Avg admissions per calendar month · all years
MONTHLY
Brazil YoY Growth Rate
Year-over-year % change
GROWTH
R² Score
0.7941
Model explains 79% of variance
MAE
593
Mean Absolute Error (admissions)
MAPE
64.7%
Driven by small-value categories
Within ±20%
48.3%
30.8% within ±10%
Actual vs. Predicted
Test set · 474 records · 2023–2024
SCATTER
Reading this chart
Points along the diagonal = perfect prediction. The model tracks large-volume categories well (Ontario, Alberta). Outliers are mostly niche categories with low absolute volume — high MAPE, low strategic impact.
Feature Importance
What drives the model's predictions?
EXPLAINABILITY
Key Takeaway
Rolling 3-year average (90.1%) is by far the dominant signal — past trends are the best predictor of future volume. This validates the strategic hypothesis: historical patterns are the real intelligence asset, not one-off spikes.
Residual Distribution
Prediction error spread across test set
RESIDUALS
Distribution Reading
The vast majority of errors cluster near zero — the model is unbiased. The few large negative residuals (model over-predicted) correspond to Ontario's 2024 policy-driven drop, which no historical model can anticipate without external signals.
2025 Forecast — Top Provinces
Model prediction vs 2024 actual
FORECAST
Forecast Context
Quebec and Saskatchewan show strong projected growth for 2025 based on their 2023–2024 trajectory. Alberta stabilizes. These signals help consulting firms advise clients on where to anchor their destination province for maximum approval probability.