-
Notifications
You must be signed in to change notification settings - Fork 0
panel construction
Syed Ibrahim Omer edited this page Apr 12, 2026
·
5 revisions
The feature engineering stage that merges all data sources into a single monthly × country panel dataset ready for model training.
src/processing/build_panel.py — see Build Panel Detail for technical deep dive.
| Source | Granularity | Module |
|---|---|---|
| Visa Data | Monthly issuances by country | visa_master.parquet |
| Exchange Rates | Monthly average REER | OECD data |
| Google News | Event counts per month by country | news_embeddings_labeled/ |
| Google Trends | Monthly keyword interest | Trend Parquets |
For each of 3 core signals, 6 months of history:
| Signal | Features |
|---|---|
| Visa volume |
visa_lag_1 through visa_lag_6
|
| Exchange rate |
exchange_rate_lag_1 through exchange_rate_lag_6
|
| News events |
news_events_lag_1 through news_events_lag_6
|
Forward-looking prediction targets:
-
target_visa_lead_1throughtarget_visa_lead_6
| Metric | Value |
|---|---|
| Countries | 15 |
| Months | ~120 (2017–2025) |
| Total observations | ~1,800 |
| Features per observation | 18 lag + metadata |
| Targets per observation | 6 leads |
- Exchange rates: Forward/backward fill for gaps; zero-fill for countries without REER data (9 of 15)
- News events: Zero-fill for months with no articles
- Visa lags: Drop observations with insufficient history (first 6 months per country)
data/processed/train_panel.parquet — consumed by Training Pipeline and Inference Pipeline.
The model layer reshapes lag features into sequences:
(batch_size, 6, 3) → 6 time steps × 3 signals [visa, exchange_rate, news_events]
This is handled by build_sequential_tensors() in src/models/surge_model.py with StandardScaler normalization.
- Build Panel Detail — Technical deep dive (forward-fill, edge cases)
- Data Processing — Previous stage
- Training Pipeline — Next stage (model training)
- Lead-Lag Analysis — Statistical basis for the lag window choice
- Project Overview — Goals, research questions, methodology, and team
- Glossary — Key terms used throughout this wiki
Raw inputs that feed the prediction system.
| Page | Description |
|---|---|
| Visa Data | US Department of State visa issuance statistics (108 monthly PDFs) |
| Encounter Data | CBP Southwest border encounter statistics (FY2019–2026) |
| Google News | 170K+ news articles across 15 countries × 8 topics |
| Google Trends | Monthly search-interest time series (15 countries × 8 keywords) |
| Exchange Rates | IMF Real Effective Exchange Rate for 6 countries |
The end-to-end flow from raw data to production forecasts.
| Page | Description |
|---|---|
| Data Collection | Ingestion layer: async scraping, bounded concurrency, retry logic |
| Data Processing | PDF parsing, JSON→Parquet, encounter merging |
| NLP Enrichment | Embedding → Clustering → Labeling → Sentiment |
| Panel Construction | Feature engineering: 18 lag features, 6 lead targets |
| Training Pipeline | Out-of-time train/test split, 4 architectures |
| Inference Pipeline | Horizon-aware ensemble, production prediction flow |
Machine learning architectures and their roles in the ensemble.
| Page | Description |
|---|---|
| Random Forest | cuML GPU Random Forest — best at short horizons (Lead 1–2) |
| LSTM | MigrationLSTM — country-aware with SurgeJointLoss |
| Transformer | MigrationTransformer — best at long horizons (Lead 5–6) |
| Horizon-Aware Ensemble | Dynamic weighting: RF→short, Transformer→long |
| SurgeJointLoss | Dual-objective loss: Huber + BCE for crisis detection |
| Jina v5 Embeddings | TensorRT INT8 news article embeddings (768-dim) |
| Flan-T5 Summarization | TensorRT INT8 cluster labeling engine |
Statistical techniques driving the lead-lag and surge analysis.
| Page | Description |
|---|---|
| Lead-Lag Analysis | Pearson correlation at 0–6 month offsets |
| Surge Detection | Quantile-based and σ-threshold spike identification |
| Sentiment Analysis | Rule-based lexicon scoring for migration-relevant news |
| Event Clustering | HDBSCAN GPU clustering + LED label generation |
| Cross-Correlation Analysis | CCF analysis, VAR benchmarking, ADF stationarity tests |
| Multiple Comparison Correction | Benjamini-Hochberg FDR for 58 significant signals |
What the system discovered about migration predictability.
| Page | Description |
|---|---|
| Event-Visa Findings | News events as leading indicators (r=0.617 at 3-month lag) |
| Exchange Rate Findings | Exchange rate signals (DR r=0.498 at 2-month lag) |
| Model Performance | Ensemble results: F1=0.96 at Lead 1, F1=0.86 at Lead 6 |
Reference documentation for every src/ subpackage and key files.
| Page | Description |
|---|---|
| Main Entry Point |
src/main.py CLI: bootstrap, collect-live, sync-data |
| Collection Module |
src/collection/* — visa, encounter, news, trends, HF sync |
| Processing Module |
src/processing/* — parse, merge, build_panel, summarize |
| Analysis Module |
src/analysis/* — events, exchange_rate, trends_analysis, plots |
| Models Module |
src/models/* — surge_model, train_and_evaluate, inference |
| News Scraper | Deep dive: batch decoding, checkpoint recovery, throttling |
| PDF Parser | Deep dive: PyMuPDF table extraction, VISA_MAP normalization |
| TensorRT Engines | Deep dive: Jina-v5, Flan-T5, LED TensorRT engines |
| Build Panel Detail | Deep dive: lag/lead construction, forward-fill strategies |
| HF Sync | Deep dive: bidirectional Hugging Face Hub sync |
Compute, reproducibility, and operational details.
| Page | Description |
|---|---|
| GPU Acceleration | TensorRT INT8, cuML, CUDA streams, NVML profiling |
| Reproducibility | HF bootstrap, run.sh pipeline, dependency checking |