maisons_avec_prix.json |
house filter + cleaning described in docs/METHODOLOGIE.md §2–4 (combined Realtor + DuProprio houses with price) |
19,123 house listings |
specs.jsonl |
scripts/01_extract_specs_llm.py (GPT-4o-mini, strict JSON schema, resumable cache keyed by uid) |
structured specifications extracted from descriptions |
facade_specs.jsonl |
scripts/02_facade_vision.py (vision–language model on the façade photo) |
objective visual attributes (curb appeal, cladding, …) |
maisons_specs.json / .csv |
scripts/03_merge_llm_features.py |
listings merged with LLM features |
maisons_hedonic.csv, maisons_hedonic_v4.csv |
scripts/04_…, scripts/05_… |
final datasets with fitted price and resid_pct |
nn_robust_oos.npy, v5_blend_oos.npy |
scripts/08_…, scripts/09_… |
out-of-sample predictions (5-fold, seed 42) reused by steps 10–14 |