Tabular Classification
Scikit-learn
Joblib
English
French
wildfire
fire-weather-index
scikit-learn
logistic-regression
france
var
open-data
Instructions to use GuillaumeGay/firepredict with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Scikit-learn
How to use GuillaumeGay/firepredict with Scikit-learn:
from huggingface_hub import hf_hub_download import joblib model = joblib.load( hf_hub_download("GuillaumeGay/firepredict", "sklearn_model.joblib") ) # only load pickle files from sources you trust # read more about it here https://skops.readthedocs.io/en/stable/persistence.html - Notebooks
- Google Colab
- Kaggle
small_commune_v2, frozen 9 November 2026
Browse files- LICENSE +21 -0
- README.md +188 -0
- example.py +104 -0
- forecasts_2026.csv +0 -0
- metadata.json +153 -0
- model.joblib +3 -0
- usual_rates.csv +154 -0
LICENSE
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
MIT License
|
| 2 |
+
|
| 3 |
+
Copyright (c) 2026 Guillaume Gay
|
| 4 |
+
|
| 5 |
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
| 6 |
+
of this software and associated documentation files (the "Software"), to deal
|
| 7 |
+
in the Software without restriction, including without limitation the rights
|
| 8 |
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
| 9 |
+
copies of the Software, and to permit persons to whom the Software is
|
| 10 |
+
furnished to do so, subject to the following conditions:
|
| 11 |
+
|
| 12 |
+
The above copyright notice and this permission notice shall be included in all
|
| 13 |
+
copies or substantial portions of the Software.
|
| 14 |
+
|
| 15 |
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
| 16 |
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
| 17 |
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
| 18 |
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
| 19 |
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
| 20 |
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
| 21 |
+
SOFTWARE.
|
README.md
ADDED
|
@@ -0,0 +1,188 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
- fr
|
| 6 |
+
library_name: sklearn
|
| 7 |
+
pipeline_tag: tabular-classification
|
| 8 |
+
tags:
|
| 9 |
+
- wildfire
|
| 10 |
+
- fire-weather-index
|
| 11 |
+
- tabular-classification
|
| 12 |
+
- scikit-learn
|
| 13 |
+
- logistic-regression
|
| 14 |
+
- france
|
| 15 |
+
- var
|
| 16 |
+
- open-data
|
| 17 |
+
---
|
| 18 |
+
|
| 19 |
+
# firepredict: daily wildfire-start risk for the communes of the Var (France)
|
| 20 |
+
|
| 21 |
+
A small logistic regression that gives each of the 153 communes of the Var (83), France, a daily chance that a fire of 0.1 ha or more starts there. It is built only from open data. Code, data pipeline and full report: [github.com/guillaumegay13/firepredict](https://github.com/guillaumegay13/firepredict).
|
| 22 |
+
|
| 23 |
+
> **Research preview.** This is not an official warning tool. Fire danger warnings in France are issued by Météo-France and the prefectures. The every-year results below are exploratory (see [Limitations](#limitations)); the clean test is the 2026 one.
|
| 24 |
+
|
| 25 |
+
## What it predicts
|
| 26 |
+
|
| 27 |
+
For one commune and one target day: the chance that **at least one fire of 0.1 ha or more starts in the commune on the target day**.
|
| 28 |
+
|
| 29 |
+
- The forecast is issued on the morning of the day before the target day.
|
| 30 |
+
- It uses data up to two days before the target day (the last complete day of station data).
|
| 31 |
+
- No input uses recent fire records. BDIFF publishes a year's fires months later, so a live forecast cannot know them. Automated tests in the GitHub repository check the time rules and that no input changes when fire records change.
|
| 32 |
+
|
| 33 |
+
Version: `small_commune_v2`, frozen on 9 November 2026 (git commit `5fa2bc1`), trained on every labelled year.
|
| 34 |
+
|
| 35 |
+
## Inputs
|
| 36 |
+
|
| 37 |
+
A scikit-learn pipeline: `StandardScaler` then `LogisticRegression` (default settings, `max_iter=1000`). Five inputs, in this order:
|
| 38 |
+
|
| 39 |
+
| Input | Definition | Unit / transform | Standardised weight |
|
| 40 |
+
|---|---|---|---|
|
| 41 |
+
| `log_usual_rate` | Commune's usual fire rate: average number of days per year with at least one fire start of 0.1 ha or more, 2006–2025 without 2024 (`usual_rates.csv`) | log(rate + 0.01) | +0.589 |
|
| 42 |
+
| `log_local_fwi` | Fire Weather Index two days before the target day, from the commune's nearest Météo-France stations | log(1 + FWI), FWI has no unit | +0.858 (strongest) |
|
| 43 |
+
| `local_wb_30d` | Rain minus potential evaporation (Météo-France ETP) over the 30 days that end two days before the target day, nearest stations | mm, negative = dry | −0.562 |
|
| 44 |
+
| `orange_period` | Target day in February or March (brush burning allowed under conditions) | 0 / 1 | +0.230 |
|
| 45 |
+
| `red_period` | Target day in June to September (fire use banned) | 0 / 1 | −0.261 |
|
| 46 |
+
|
| 47 |
+
A missing input is replaced with its training median (`fill_values` in `metadata.json`).
|
| 48 |
+
|
| 49 |
+
The FWI here is an approximation computed from daily station values (daily maximum temperature, minimum humidity, mean wind, daily rain), not the official noon-based product.
|
| 50 |
+
|
| 51 |
+
## How to use
|
| 52 |
+
|
| 53 |
+
```bash
|
| 54 |
+
uv run --with scikit-learn==1.9.1 --with joblib --with pandas python example.py
|
| 55 |
+
```
|
| 56 |
+
|
| 57 |
+
`example.py` checks the SHA-256 fingerprints, loads the model and computes the chance for example rows. The pipeline was saved with **scikit-learn 1.9.1**; older versions (for example 1.7.2) fail to load it. Joblib files are pickles: load only files whose fingerprint you trust (`model_sha256` in `metadata.json`).
|
| 58 |
+
|
| 59 |
+
```python
|
| 60 |
+
import joblib, json, numpy as np, pandas as pd
|
| 61 |
+
|
| 62 |
+
model = joblib.load("model.joblib")
|
| 63 |
+
meta = json.load(open("metadata.json"))
|
| 64 |
+
rates = pd.read_csv("usual_rates.csv", dtype={"insee": str}).set_index("insee").usual_rate
|
| 65 |
+
|
| 66 |
+
X = pd.DataFrame([{
|
| 67 |
+
"log_usual_rate": np.log(rates["83137"] + 0.01), # Toulon
|
| 68 |
+
"log_local_fwi": np.log1p(35.0), # FWI two days before the target day
|
| 69 |
+
"local_wb_30d": -140.0, # mm over 30 days
|
| 70 |
+
"orange_period": 0,
|
| 71 |
+
"red_period": 1, # target day in July
|
| 72 |
+
}])[meta["inputs"]].fillna(pd.Series(meta["fill_values"]))
|
| 73 |
+
print(model.predict_proba(X)[0, 1]) # about 0.0034
|
| 74 |
+
```
|
| 75 |
+
|
| 76 |
+
To build the inputs from raw data (weather download, station matching, FWI), use the GitHub repository: `firepredict.model_communes.build_panel` and `firepredict.freeze.predict`, or `uv run python -m firepredict.pipeline forecast` for today's forecast.
|
| 77 |
+
|
| 78 |
+
## Files
|
| 79 |
+
|
| 80 |
+
| File | Content |
|
| 81 |
+
|---|---|
|
| 82 |
+
| `model.joblib` | The fitted pipeline (SHA-256 `a1e64b40ed3aa7c274d4b2ec01f24e93a80beb976c3e03ce6d2a3e837056b243`) |
|
| 83 |
+
| `usual_rates.csv` | Usual fire rate per commune (INSEE code), used for forecasts |
|
| 84 |
+
| `metadata.json` | Inputs, weights, fill values, training years and rows, raw-data fingerprints, file fingerprints |
|
| 85 |
+
| `forecasts_2026.csv` | Retrospective 2026 test forecasts: `target_day`, `insee`, `chance` (47,889 rows) |
|
| 86 |
+
| `example.py` | Load the model and compute a chance |
|
| 87 |
+
| `LICENSE` | MIT licence of the code and model |
|
| 88 |
+
|
| 89 |
+
## Training data
|
| 90 |
+
|
| 91 |
+
- **Fire records:** BDIFF (Base de Données sur les Incendies de Forêts en France), Var, 2006–2025 without 2024. Only fires of 0.1 ha or more are counted.
|
| 92 |
+
- **Weather:** Météo-France daily climatological data, all Var stations.
|
| 93 |
+
- **Burning calendar:** Var prefectoral orders on fire use of 16 May 2013 and 2 November 2025.
|
| 94 |
+
|
| 95 |
+
1,052,640 commune-days, of which 902 positive commune-days (from 910 fires). A positive commune-day is a day on which at least one fire of 0.1 ha or more started in the commune.
|
| 96 |
+
|
| 97 |
+
## Evaluation
|
| 98 |
+
|
| 99 |
+
Leave-one-year-out over 19 years (2006–2025 without 2024): each year is hidden in turn, the model learns on the other years, then forecasts the hidden year. The unit is a positive commune-day (902 in total). Ranges are 95% ranges from resampling years.
|
| 100 |
+
|
| 101 |
+
**Against two references:**
|
| 102 |
+
|
| 103 |
+
| Reference | ROC AUC (reference → model) | Years better | 15 communes/day: caught | 10% of commune-days: caught |
|
| 104 |
+
|---|---|---|---|---|
|
| 105 |
+
| FWI alone (Var) | 0.718 → 0.782 | 19 / 19 | 85 → 213, +151% (+104% to +221%) | 230 → 354, +54% (+33% to +74%) |
|
| 106 |
+
| Local FWI alone | 0.728 → 0.782 | 18 / 19 | 123 → 213, +73% (+46% to +111%) | 256 → 354, +38% (+20% to +58%) |
|
| 107 |
+
| FWI × commune history | 0.757 → 0.782 | 16 / 19 (p < 0.001) | 222 → 213, −4% (−10% to +2%) | 307 → 354, +15% (+2% to +30%) |
|
| 108 |
+
|
| 109 |
+
- **Against the FWI alone**, the model is clearly better.
|
| 110 |
+
- **Against the FWI multiplied by each commune's usual fire rate**, the model ranks commune-days better, but the gain in fires caught is about zero at practical warning levels (15 communes a day: −4%, −10% to +2%).
|
| 111 |
+
- **On its very top warnings the model does worse than FWI × history:** 59 against 76 at the 1% most alarming commune-days. Its pooled PR AUC is also lower (0.0030 against 0.0034).
|
| 112 |
+
|
| 113 |
+
**What each input adds** (nested, ROC AUC):
|
| 114 |
+
|
| 115 |
+
| Inputs | ROC AUC |
|
| 116 |
+
|---|---|
|
| 117 |
+
| Commune history + local FWI | 0.767 |
|
| 118 |
+
| + burning calendar | 0.775 |
|
| 119 |
+
| + 30-day dryness (this model) | 0.782 |
|
| 120 |
+
| + buildings and roads in wildland (IGN BD TOPO) | 0.781 |
|
| 121 |
+
|
| 122 |
+
Buildings and roads near vegetation add nothing once each commune's fire history is known, so they are not in the model.
|
| 123 |
+
|
| 124 |
+
**Calibration** holds in the bands used by the map: commune-days predicted at 0.1–0.3% had a mean prediction of 0.167% and an observed rate of 0.169%; those at 0.3–1%, 0.38% predicted against 0.45% observed. Above 1% there are only 17 commune-days, too few to check.
|
| 125 |
+
|
| 126 |
+
**Robustness** (model against FWI × commune history):
|
| 127 |
+
|
| 128 |
+
- Without 2023: ROC AUC 0.787 against 0.761 (better in 16 of 18 years).
|
| 129 |
+
- Fires of 1 ha or more: ROC AUC 0.780 against 0.720 (better in 16 of 19 years).
|
| 130 |
+
|
| 131 |
+
**Location within a day is hard to improve.** None of the fixed rankings tested puts more than about 25% of fire starts in the day's 15 highest-ranked communes. This is what the rankings tested reach, not a proven ceiling.
|
| 132 |
+
|
| 133 |
+
Larger models (CatBoost, the Timestral time-series model) were tested in the project too. They do no better and are less consistent from year to year.
|
| 134 |
+
|
| 135 |
+
These results were corrected after an independent review of the code and claims on 9 November 2026 (inputs built from recent fire records were removed, two small leaks were closed, wording was corrected). See the "Update after an independent review" section of [`reports/phase1_results.md`](https://github.com/guillaumegay13/firepredict/blob/main/reports/phase1_results.md).
|
| 136 |
+
|
| 137 |
+
## 2026 retrospective test and live archive
|
| 138 |
+
|
| 139 |
+
The inputs were chosen after looking at results across all years, so the scores above are somewhat optimistic. Two cleaner tests are running:
|
| 140 |
+
|
| 141 |
+
- **Retrospective 2026 test, blind to the outcomes.** On 9 November 2026, before any 2026 fire record was published, the model was frozen and its forecasts for 1 January – 9 November 2026 were saved in `forecasts_2026.csv` (SHA-256 `2be6091ab25539dc47bb751822d2a9b2b48f21218f8238bf0be1e1653dc8c6d6`). They were all computed on that day, from the weather available then. This is a retrospective test, not a live trial. When BDIFF publishes the 2026 fires, the forecasts will be scored as they are.
|
| 142 |
+
- **Live archive.** From 9 November 2026, each day's forecast from the frozen model is saved in the [`forecasts/`](https://github.com/guillaumegay13/firepredict/tree/main/forecasts) folder of the GitHub repository and committed the same day, so the git history dates it.
|
| 143 |
+
|
| 144 |
+
## Limitations
|
| 145 |
+
|
| 146 |
+
- **Exploratory results.** The inputs and the 30-day window were chosen after looking at the same years used for evaluation, so p-values and ranges are optimistic. The ranges are conditional on the chosen model.
|
| 147 |
+
- **Approximate FWI.** Computed from daily station values, not the official noon-based FWI or Météo-France's forecasts.
|
| 148 |
+
- **One department.** The model is trained and tested on the Var only. It has not been checked on other departments.
|
| 149 |
+
- **Label issues.** 2024 is excluded: BDIFF holds only 3 Var fires for 2024, against 50–380 in other years. In 2023 BDIFF started recording many more tiny fires, so only fires of 0.1 ha or more are counted.
|
| 150 |
+
- **Commune-level locations only.** BDIFF gives the commune of each fire, not its position.
|
| 151 |
+
- **Burning calendar.** The 2013 and 2025 rules are applied to all years. The 2006–2012 rules and the prefect's extra red days are not included.
|
| 152 |
+
- **Past features use later knowledge.** Weather stations were chosen by their coverage over 2006–2026, and the leave-one-year-out test trains on years after the one it scores.
|
| 153 |
+
- **Small chances.** Most predicted chances are below 1% per commune-day; they are useful to rank communes and days, not as yes/no forecasts.
|
| 154 |
+
|
| 155 |
+
## Data and credits
|
| 156 |
+
|
| 157 |
+
The model was trained on the data below. This repository does not contain BDIFF or Météo-France raw data; the code in the GitHub repository downloads it from the official sites. `usual_rates.csv` holds yearly averages per commune derived from BDIFF.
|
| 158 |
+
|
| 159 |
+
| Data | Provider | Licence |
|
| 160 |
+
|---|---|---|
|
| 161 |
+
| Fire records 2006–2025 (BDIFF) | Ministère de l'Agriculture, [bdiff.agriculture.gouv.fr](https://bdiff.agriculture.gouv.fr) | see below |
|
| 162 |
+
| Daily weather, all Var stations | Météo-France, via [data.gouv.fr](https://www.data.gouv.fr/datasets/donnees-climatologiques-de-base-quotidiennes) | Licence Ouverte / Etalab 2.0 |
|
| 163 |
+
| Burning calendar | Var prefectoral orders of 16 May 2013 and 2 November 2025 | public acts |
|
| 164 |
+
| Commune boundaries (to match each commune to its nearest stations) | geo.api.gouv.fr (IGN, INSEE) | Licence Ouverte / Etalab 2.0 |
|
| 165 |
+
|
| 166 |
+
The wider project (GitHub repository) also uses IGN BD TOPO, DDTM du Var burned-area outlines, Corine Land Cover 2018 and school holiday data; none of them is an input of this model.
|
| 167 |
+
|
| 168 |
+
**BDIFF.** Source : Base de Données sur les Incendies de Forêts en France, [bdiff.agriculture.gouv.fr](https://bdiff.agriculture.gouv.fr). Droits de reproduction réservés et strictement limités. BDIFF data may be reused for personal, associative or professional purposes, unchanged, with this source cited, and may not be sold or rented.
|
| 169 |
+
|
| 170 |
+
## Licence
|
| 171 |
+
|
| 172 |
+
Code and model under the MIT licence (see `LICENSE`). Data under the licences above.
|
| 173 |
+
|
| 174 |
+
## Citation
|
| 175 |
+
|
| 176 |
+
```bibtex
|
| 177 |
+
@software{gay2026firepredict,
|
| 178 |
+
author = {Gay, Guillaume},
|
| 179 |
+
title = {firepredict: daily wildfire-start risk for the communes of the Var, France, from open data},
|
| 180 |
+
year = {2026},
|
| 181 |
+
url = {https://github.com/guillaumegay13/firepredict},
|
| 182 |
+
note = {Model small\_commune\_v2, frozen 9 November 2026}
|
| 183 |
+
}
|
| 184 |
+
```
|
| 185 |
+
|
| 186 |
+
## Contact
|
| 187 |
+
|
| 188 |
+
Guillaume Gay, guillaume.gay@protonmail.com. Feedback from fire services, researchers and local authorities is very welcome.
|
example.py
ADDED
|
@@ -0,0 +1,104 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Load the frozen firepredict model and compute a chance for one commune-day.
|
| 2 |
+
|
| 3 |
+
Run from this folder:
|
| 4 |
+
uv run --with scikit-learn==1.9.1 --with joblib --with pandas python example.py
|
| 5 |
+
|
| 6 |
+
The model answers: what is the chance that at least one fire of 0.1 ha or more
|
| 7 |
+
starts in this Var commune on the target day? The forecast is issued on the
|
| 8 |
+
morning of the day before the target day and uses weather up to two days
|
| 9 |
+
before the target day.
|
| 10 |
+
|
| 11 |
+
The five inputs (one row per commune and target day):
|
| 12 |
+
|
| 13 |
+
log_usual_rate log(usual_rate + 0.01). usual_rate is the commune's average
|
| 14 |
+
number of days per year with at least one fire start of
|
| 15 |
+
0.1 ha or more (BDIFF, 2006-2025 without 2024). Take it from
|
| 16 |
+
usual_rates.csv by INSEE code.
|
| 17 |
+
log_local_fwi log(1 + FWI). FWI is the Fire Weather Index (no unit) of the
|
| 18 |
+
day two days before the target day, computed from the
|
| 19 |
+
commune's nearest Météo-France stations with daily values
|
| 20 |
+
(maximum temperature, minimum humidity, mean wind, daily
|
| 21 |
+
rain). It is an approximation, not the official FWI product.
|
| 22 |
+
local_wb_30d Rain minus potential evaporation (Météo-France ETP), in mm,
|
| 23 |
+
summed over the 30 days that end two days before the target
|
| 24 |
+
day, at the commune's nearest stations. Negative = dry.
|
| 25 |
+
orange_period 1 if the target day is in February or March (brush burning
|
| 26 |
+
allowed under conditions, Var prefectoral orders), else 0.
|
| 27 |
+
red_period 1 if the target day is in June to September (fire use
|
| 28 |
+
banned), else 0.
|
| 29 |
+
|
| 30 |
+
A missing input is replaced with the training median stored in
|
| 31 |
+
metadata.json ("fill_values"), as the frozen pipeline does.
|
| 32 |
+
|
| 33 |
+
To build these inputs from raw data (weather download, station matching,
|
| 34 |
+
FWI), use the code at https://github.com/guillaumegay13/firepredict
|
| 35 |
+
(firepredict.model_communes.build_panel and firepredict.freeze.predict).
|
| 36 |
+
|
| 37 |
+
This is a research preview, not an official warning tool. Fire danger
|
| 38 |
+
warnings in France come from Météo-France and the prefectures.
|
| 39 |
+
"""
|
| 40 |
+
|
| 41 |
+
import hashlib
|
| 42 |
+
import json
|
| 43 |
+
from pathlib import Path
|
| 44 |
+
|
| 45 |
+
import joblib
|
| 46 |
+
import numpy as np
|
| 47 |
+
import pandas as pd
|
| 48 |
+
import sklearn
|
| 49 |
+
|
| 50 |
+
HERE = Path(__file__).parent
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
def sha256(path: Path) -> str:
|
| 54 |
+
return hashlib.sha256(path.read_bytes()).hexdigest()
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
# The pipeline was saved with scikit-learn 1.9.1; older versions fail to load it.
|
| 58 |
+
if sklearn.__version__ != "1.9.1":
|
| 59 |
+
print(f"warning: scikit-learn {sklearn.__version__} found, the model was saved with 1.9.1")
|
| 60 |
+
|
| 61 |
+
metadata = json.loads((HERE / "metadata.json").read_text())
|
| 62 |
+
|
| 63 |
+
# joblib files are pickles: only load files whose fingerprint you trust.
|
| 64 |
+
assert sha256(HERE / "model.joblib") == metadata["model_sha256"], "model.joblib does not match metadata.json"
|
| 65 |
+
assert sha256(HERE / "forecasts_2026.csv") == metadata["forecasts_2026"]["sha256"], "forecasts_2026.csv changed"
|
| 66 |
+
|
| 67 |
+
model = joblib.load(HERE / "model.joblib")
|
| 68 |
+
rates = pd.read_csv(HERE / "usual_rates.csv", dtype={"insee": str}).set_index("insee").usual_rate
|
| 69 |
+
|
| 70 |
+
|
| 71 |
+
def inputs(insee: str, fwi: float, wb_30d_mm: float, target_month: int) -> pd.DataFrame:
|
| 72 |
+
"""One row of the five model inputs, in the order the model expects."""
|
| 73 |
+
row = {
|
| 74 |
+
"log_usual_rate": np.log(rates.get(insee, np.nan) + 0.01),
|
| 75 |
+
"log_local_fwi": np.log1p(fwi),
|
| 76 |
+
"local_wb_30d": wb_30d_mm,
|
| 77 |
+
"orange_period": int(target_month in (2, 3)),
|
| 78 |
+
"red_period": int(target_month in (6, 7, 8, 9)),
|
| 79 |
+
}
|
| 80 |
+
X = pd.DataFrame([row])[metadata["inputs"]]
|
| 81 |
+
return X.fillna(pd.Series(metadata["fill_values"]))
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
examples = [
|
| 85 |
+
# (description, INSEE code, FWI two days before, 30-day rain - evaporation in mm, month of target day)
|
| 86 |
+
("Toulon, dry July day", "83137", 35.0, -140.0, 7),
|
| 87 |
+
("Toulon, February day", "83137", 8.0, -20.0, 2),
|
| 88 |
+
("Toulon, wet November day", "83137", 2.0, 60.0, 11),
|
| 89 |
+
]
|
| 90 |
+
|
| 91 |
+
print(f"model {metadata['version']}, frozen {metadata['frozen_at']}, inputs {metadata['inputs']}")
|
| 92 |
+
print(f"usual rate of Toulon (83137): {rates['83137']:.3f} days with a fire start per year\n")
|
| 93 |
+
for label, insee, fwi, wb, month in examples:
|
| 94 |
+
X = inputs(insee, fwi, wb, month)
|
| 95 |
+
chance = model.predict_proba(X)[0, 1]
|
| 96 |
+
# Same result by hand: standardise, weighted sum, logistic function.
|
| 97 |
+
scaler, logit = model[0], model[-1]
|
| 98 |
+
z = ((X.to_numpy()[0] - scaler.mean_) / scaler.scale_) @ logit.coef_[0] + logit.intercept_[0]
|
| 99 |
+
assert np.isclose(chance, 1 / (1 + np.exp(-z)))
|
| 100 |
+
print(f"{label:28s} {X.round(3).to_dict('records')[0]}")
|
| 101 |
+
print(f"{'':28s} chance of a fire start: {chance:.4%}\n")
|
| 102 |
+
|
| 103 |
+
forecasts = pd.read_csv(HERE / "forecasts_2026.csv", dtype={"insee": str})
|
| 104 |
+
print(f"forecasts_2026.csv: {len(forecasts)} rows, {forecasts.target_day.min()} to {forecasts.target_day.max()}")
|
forecasts_2026.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
metadata.json
ADDED
|
@@ -0,0 +1,153 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"version": "small_commune_v2",
|
| 3 |
+
"frozen_at": "2026-11-09T12:35:04+00:00",
|
| 4 |
+
"git_commit": "5fa2bc109a9eccab69a1f26349baec04b89ad6d9",
|
| 5 |
+
"question": "chance that at least one fire >= 0.1 ha starts in the commune on the target day; issued the morning before, from data up to two days before the target day",
|
| 6 |
+
"inputs": [
|
| 7 |
+
"log_usual_rate",
|
| 8 |
+
"log_local_fwi",
|
| 9 |
+
"local_wb_30d",
|
| 10 |
+
"orange_period",
|
| 11 |
+
"red_period"
|
| 12 |
+
],
|
| 13 |
+
"weights_standardised": {
|
| 14 |
+
"log_usual_rate": 0.5893,
|
| 15 |
+
"log_local_fwi": 0.8582,
|
| 16 |
+
"local_wb_30d": -0.5619,
|
| 17 |
+
"orange_period": 0.2303,
|
| 18 |
+
"red_period": -0.2605
|
| 19 |
+
},
|
| 20 |
+
"fill_values": {
|
| 21 |
+
"log_usual_rate": -1.460061,
|
| 22 |
+
"log_local_fwi": 2.703165,
|
| 23 |
+
"local_wb_30d": -34.599998,
|
| 24 |
+
"orange_period": 0.0,
|
| 25 |
+
"red_period": 0.0
|
| 26 |
+
},
|
| 27 |
+
"training_years": [
|
| 28 |
+
2006,
|
| 29 |
+
2007,
|
| 30 |
+
2008,
|
| 31 |
+
2009,
|
| 32 |
+
2010,
|
| 33 |
+
2011,
|
| 34 |
+
2012,
|
| 35 |
+
2013,
|
| 36 |
+
2014,
|
| 37 |
+
2015,
|
| 38 |
+
2016,
|
| 39 |
+
2017,
|
| 40 |
+
2018,
|
| 41 |
+
2019,
|
| 42 |
+
2020,
|
| 43 |
+
2021,
|
| 44 |
+
2022,
|
| 45 |
+
2023,
|
| 46 |
+
2025
|
| 47 |
+
],
|
| 48 |
+
"training_rows": 1052640,
|
| 49 |
+
"training_positive_commune_days": 902,
|
| 50 |
+
"bdtopo_edition": "BDTOPO_3-5_TOUSTHEMES_GPKG_LAMB93_D083_2026-09-15",
|
| 51 |
+
"raw_data": {
|
| 52 |
+
"bdiff/bdiff_83_2006-2025.zip": {
|
| 53 |
+
"sha256": "a215fc8bb6c2a2eeca482bc705de2fc5cc3c87b2634a6e86934ffcd0e7ee4103",
|
| 54 |
+
"downloaded_at": "2026-10-27T08:10:12+00:00"
|
| 55 |
+
},
|
| 56 |
+
"meteofrance/Q_83_previous-1950-2024_RR-T-Vent.csv.gz": {
|
| 57 |
+
"sha256": "3c0fe43a70278c6a72499802499f15c0e6bfd0cc74db8674eb8bd3961c48c97a",
|
| 58 |
+
"downloaded_at": "2026-10-27T08:10:14+00:00"
|
| 59 |
+
},
|
| 60 |
+
"meteofrance/Q_83_previous-1950-2024_autres-parametres.csv.gz": {
|
| 61 |
+
"sha256": "3d3d672ae7594a321dec071b3dc3567ec397a4c7f1312f880359666dfd59ba6f",
|
| 62 |
+
"downloaded_at": "2026-10-27T08:10:15+00:00"
|
| 63 |
+
},
|
| 64 |
+
"meteofrance/Q_83_latest-2025-2026_RR-T-Vent.csv.gz": {
|
| 65 |
+
"sha256": "1910c95009245f0c79ab288db0141c3f9f187f16c250ef3fb601e03e98e02194",
|
| 66 |
+
"downloaded_at": "2026-11-09T12:29:13+00:00"
|
| 67 |
+
},
|
| 68 |
+
"meteofrance/Q_83_latest-2025-2026_autres-parametres.csv.gz": {
|
| 69 |
+
"sha256": "2f9bdfc107f5984239a08f331ddd5b760ff4858246fbe43de40fc2e33f74c061",
|
| 70 |
+
"downloaded_at": "2026-11-09T12:29:14+00:00"
|
| 71 |
+
},
|
| 72 |
+
"ddtm83/contours_feux_var_1958-2022.zip": {
|
| 73 |
+
"sha256": "c4c10e0abaaa763807104d5bb9690702115a01842e9589881b7630901daa91c2",
|
| 74 |
+
"downloaded_at": "2026-10-27T08:10:33+00:00"
|
| 75 |
+
},
|
| 76 |
+
"ddtm83/contours_feux_var_2008.zip": {
|
| 77 |
+
"sha256": "2ec3c7903ebd3347c0afe2ccda2248b8895c013f166d15aba426719b88690bfd",
|
| 78 |
+
"downloaded_at": "2026-10-27T08:10:35+00:00"
|
| 79 |
+
},
|
| 80 |
+
"ddtm83/contours_feux_var_2010.zip": {
|
| 81 |
+
"sha256": "cc2d8f32dc504617fc44c8ad1bf58c0c5d6fc63d2c32d01641cd4bb3e3afffff",
|
| 82 |
+
"downloaded_at": "2026-10-27T08:10:36+00:00"
|
| 83 |
+
},
|
| 84 |
+
"ddtm83/contours_feux_var_2006.zip": {
|
| 85 |
+
"sha256": "f785f1080380f472bd2eaa5d558052bc22328ff7b2a857d50a959cfc0e4793ae",
|
| 86 |
+
"downloaded_at": "2026-10-27T08:10:37+00:00"
|
| 87 |
+
},
|
| 88 |
+
"ddtm83/contours_feux_var_2009.zip": {
|
| 89 |
+
"sha256": "7fe4864ef1d3577e78824fe23658f8bbda00f92d73efe1b1323b53f0a9f8a24a",
|
| 90 |
+
"downloaded_at": "2026-10-27T08:10:38+00:00"
|
| 91 |
+
},
|
| 92 |
+
"ddtm83/contours_feux_var_2007.zip": {
|
| 93 |
+
"sha256": "e3e9a37bc88e905fe9c80387c93e5c19b7081fd671b771082b3abe838f71144f",
|
| 94 |
+
"downloaded_at": "2026-10-27T08:10:40+00:00"
|
| 95 |
+
},
|
| 96 |
+
"ddtm83/contours_feux_var_2016.zip": {
|
| 97 |
+
"sha256": "8381568f681d081ae1f0f6153c3e6bbab41be7c3799155d50376e6aeb6a3c0e9",
|
| 98 |
+
"downloaded_at": "2026-10-27T08:10:41+00:00"
|
| 99 |
+
},
|
| 100 |
+
"ddtm83/contours_feux_var_2019.zip": {
|
| 101 |
+
"sha256": "fe1bcbb3fe2c282ce6a8a57d2b7f3ae6a16336340a2f507d37a8f26373073066",
|
| 102 |
+
"downloaded_at": "2026-10-27T08:10:42+00:00"
|
| 103 |
+
},
|
| 104 |
+
"ddtm83/contours_feux_var_2017.zip": {
|
| 105 |
+
"sha256": "4bde103598e51621637c8c69f5f3fa04f2e32e19d81e164f302cd8a40823143e",
|
| 106 |
+
"downloaded_at": "2026-10-27T08:10:43+00:00"
|
| 107 |
+
},
|
| 108 |
+
"ddtm83/contours_feux_var_2005.zip": {
|
| 109 |
+
"sha256": "e1b0edf97218a218ed083adb3d144470cf0fa7a21eabec48c6923c7b7e5c1cf2",
|
| 110 |
+
"downloaded_at": "2026-10-27T08:10:43+00:00"
|
| 111 |
+
},
|
| 112 |
+
"ddtm83/contours_feux_var_2004.zip": {
|
| 113 |
+
"sha256": "e558f8508f0d64eb4ffa27af6ba3b808ee5c22d0430ca6cb3fe1e21ee73a9689",
|
| 114 |
+
"downloaded_at": "2026-10-27T08:10:44+00:00"
|
| 115 |
+
},
|
| 116 |
+
"ddtm83/contours_feux_var_2003.zip": {
|
| 117 |
+
"sha256": "95c258c00fce3e11cdbfec8c6e9d3dabceb1014b580ef7424560f887ee6a4224",
|
| 118 |
+
"downloaded_at": "2026-10-27T08:10:45+00:00"
|
| 119 |
+
},
|
| 120 |
+
"ddtm83/contours_feux_var_2015.zip": {
|
| 121 |
+
"sha256": "8510807ac795ffdf36900d8df3b5127b724dc0a2d4af9b3d5d7528ea67909419",
|
| 122 |
+
"downloaded_at": "2026-10-27T08:10:46+00:00"
|
| 123 |
+
},
|
| 124 |
+
"calendar/vacances_scolaires.csv": {
|
| 125 |
+
"sha256": "cce6ac7d78721c424122b7eac868f6f68199efeda0dcaf9a85687d7c83bc08be",
|
| 126 |
+
"downloaded_at": "2026-10-27T08:34:33+00:00"
|
| 127 |
+
},
|
| 128 |
+
"geography/communes_83.geojson": {
|
| 129 |
+
"sha256": "82270e1304eb96b5204da728804b3fe7f8c27e8d6fd00264ed40fc36ab12b4da",
|
| 130 |
+
"downloaded_at": "2026-10-27T09:56:17+00:00"
|
| 131 |
+
},
|
| 132 |
+
"geography/corine_2018_83.geojson": {
|
| 133 |
+
"sha256": "d43ef795cee37921767104da118077d90ebbef2ac41654dc7dba3f8f300cd4e0",
|
| 134 |
+
"downloaded_at": "2026-10-27T09:56:30+00:00"
|
| 135 |
+
},
|
| 136 |
+
"bdtopo/BDTOPO_3-5_TOUSTHEMES_GPKG_LAMB93_D083_2026-09-15.7z": {
|
| 137 |
+
"sha256": "2617340d9131f0bbcc94a2fc609d8e0242d6983d1e40388fb6e06a370c955336",
|
| 138 |
+
"downloaded_at": "2026-10-28T14:20:01+00:00"
|
| 139 |
+
},
|
| 140 |
+
"bdtopo/BDTOPO_3-5_TOUSTHEMES_GPKG_LAMB93_D083_2026-09-15.gpkg": {
|
| 141 |
+
"sha256": "1b8f72705af10be7f60bf01853da66c6ef22ee43ebe755ac1204e1c1f9d1381b",
|
| 142 |
+
"downloaded_at": "2026-10-28T14:20:25+00:00"
|
| 143 |
+
}
|
| 144 |
+
},
|
| 145 |
+
"model_sha256": "a1e64b40ed3aa7c274d4b2ec01f24e93a80beb976c3e03ce6d2a3e837056b243",
|
| 146 |
+
"forecasts_2026": {
|
| 147 |
+
"kind": "outcome-blinded retrospective test",
|
| 148 |
+
"first_day": "2026-01-01",
|
| 149 |
+
"last_day": "2026-11-09",
|
| 150 |
+
"rows": 47889,
|
| 151 |
+
"sha256": "2be6091ab25539dc47bb751822d2a9b2b48f21218f8238bf0be1e1653dc8c6d6"
|
| 152 |
+
}
|
| 153 |
+
}
|
model.joblib
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1e64b40ed3aa7c274d4b2ec01f24e93a80beb976c3e03ce6d2a3e837056b243
|
| 3 |
+
size 1777
|
usual_rates.csv
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
insee,usual_rate
|
| 2 |
+
83001,0.21052631578947367
|
| 3 |
+
83002,0.3684210526315789
|
| 4 |
+
83003,0.5263157894736842
|
| 5 |
+
83004,0.7368421052631579
|
| 6 |
+
83005,0.05263157894736842
|
| 7 |
+
83006,0.15789473684210525
|
| 8 |
+
83007,0.5789473684210527
|
| 9 |
+
83008,0.0
|
| 10 |
+
83009,0.2631578947368421
|
| 11 |
+
83010,0.21052631578947367
|
| 12 |
+
83011,0.3157894736842105
|
| 13 |
+
83012,0.631578947368421
|
| 14 |
+
83013,0.0
|
| 15 |
+
83014,0.0
|
| 16 |
+
83015,0.2631578947368421
|
| 17 |
+
83016,0.2631578947368421
|
| 18 |
+
83017,0.2631578947368421
|
| 19 |
+
83018,0.21052631578947367
|
| 20 |
+
83019,0.47368421052631576
|
| 21 |
+
83020,0.10526315789473684
|
| 22 |
+
83021,0.21052631578947367
|
| 23 |
+
83022,0.10526315789473684
|
| 24 |
+
83023,0.5263157894736842
|
| 25 |
+
83025,0.10526315789473684
|
| 26 |
+
83026,0.47368421052631576
|
| 27 |
+
83027,0.3157894736842105
|
| 28 |
+
83028,0.42105263157894735
|
| 29 |
+
83029,0.2631578947368421
|
| 30 |
+
83030,0.21052631578947367
|
| 31 |
+
83031,0.5263157894736842
|
| 32 |
+
83032,0.21052631578947367
|
| 33 |
+
83033,0.47368421052631576
|
| 34 |
+
83034,0.47368421052631576
|
| 35 |
+
83035,0.3157894736842105
|
| 36 |
+
83036,0.15789473684210525
|
| 37 |
+
83037,0.10526315789473684
|
| 38 |
+
83038,0.05263157894736842
|
| 39 |
+
83039,0.05263157894736842
|
| 40 |
+
83040,0.10526315789473684
|
| 41 |
+
83041,0.42105263157894735
|
| 42 |
+
83042,0.10526315789473684
|
| 43 |
+
83043,0.5263157894736842
|
| 44 |
+
83044,0.15789473684210525
|
| 45 |
+
83045,0.3684210526315789
|
| 46 |
+
83046,0.42105263157894735
|
| 47 |
+
83047,0.21052631578947367
|
| 48 |
+
83048,0.21052631578947367
|
| 49 |
+
83049,0.21052631578947367
|
| 50 |
+
83050,1.5263157894736843
|
| 51 |
+
83051,0.8421052631578947
|
| 52 |
+
83052,0.15789473684210525
|
| 53 |
+
83053,0.21052631578947367
|
| 54 |
+
83054,0.10526315789473684
|
| 55 |
+
83055,0.3157894736842105
|
| 56 |
+
83056,0.5789473684210527
|
| 57 |
+
83057,0.3157894736842105
|
| 58 |
+
83058,0.6842105263157895
|
| 59 |
+
83059,0.10526315789473684
|
| 60 |
+
83060,0.15789473684210525
|
| 61 |
+
83061,1.0
|
| 62 |
+
83062,0.21052631578947367
|
| 63 |
+
83063,0.5263157894736842
|
| 64 |
+
83064,0.15789473684210525
|
| 65 |
+
83065,0.42105263157894735
|
| 66 |
+
83066,0.42105263157894735
|
| 67 |
+
83067,0.2631578947368421
|
| 68 |
+
83068,0.5789473684210527
|
| 69 |
+
83069,0.9473684210526315
|
| 70 |
+
83070,0.3157894736842105
|
| 71 |
+
83071,0.3684210526315789
|
| 72 |
+
83072,0.9473684210526315
|
| 73 |
+
83073,0.6842105263157895
|
| 74 |
+
83074,0.05263157894736842
|
| 75 |
+
83075,0.21052631578947367
|
| 76 |
+
83076,0.15789473684210525
|
| 77 |
+
83077,0.15789473684210525
|
| 78 |
+
83078,0.05263157894736842
|
| 79 |
+
83079,0.10526315789473684
|
| 80 |
+
83080,0.2631578947368421
|
| 81 |
+
83081,0.2631578947368421
|
| 82 |
+
83082,0.2631578947368421
|
| 83 |
+
83083,0.3157894736842105
|
| 84 |
+
83084,0.05263157894736842
|
| 85 |
+
83085,0.42105263157894735
|
| 86 |
+
83086,0.6842105263157895
|
| 87 |
+
83087,0.42105263157894735
|
| 88 |
+
83088,0.21052631578947367
|
| 89 |
+
83089,0.10526315789473684
|
| 90 |
+
83090,0.47368421052631576
|
| 91 |
+
83091,0.2631578947368421
|
| 92 |
+
83092,0.3157894736842105
|
| 93 |
+
83093,0.0
|
| 94 |
+
83094,0.15789473684210525
|
| 95 |
+
83095,0.3684210526315789
|
| 96 |
+
83096,0.3157894736842105
|
| 97 |
+
83097,0.21052631578947367
|
| 98 |
+
83098,0.05263157894736842
|
| 99 |
+
83099,0.42105263157894735
|
| 100 |
+
83100,0.3157894736842105
|
| 101 |
+
83101,0.631578947368421
|
| 102 |
+
83102,0.10526315789473684
|
| 103 |
+
83103,0.15789473684210525
|
| 104 |
+
83104,0.21052631578947367
|
| 105 |
+
83105,0.0
|
| 106 |
+
83106,0.15789473684210525
|
| 107 |
+
83107,0.7368421052631579
|
| 108 |
+
83108,0.10526315789473684
|
| 109 |
+
83109,0.15789473684210525
|
| 110 |
+
83110,0.05263157894736842
|
| 111 |
+
83111,0.10526315789473684
|
| 112 |
+
83112,0.15789473684210525
|
| 113 |
+
83113,0.15789473684210525
|
| 114 |
+
83114,0.21052631578947367
|
| 115 |
+
83115,0.15789473684210525
|
| 116 |
+
83116,0.2631578947368421
|
| 117 |
+
83117,0.10526315789473684
|
| 118 |
+
83118,0.3157894736842105
|
| 119 |
+
83119,0.10526315789473684
|
| 120 |
+
83120,0.3684210526315789
|
| 121 |
+
83121,0.47368421052631576
|
| 122 |
+
83122,0.0
|
| 123 |
+
83123,0.2631578947368421
|
| 124 |
+
83124,1.0526315789473684
|
| 125 |
+
83125,0.10526315789473684
|
| 126 |
+
83126,0.42105263157894735
|
| 127 |
+
83127,0.42105263157894735
|
| 128 |
+
83128,0.0
|
| 129 |
+
83129,0.21052631578947367
|
| 130 |
+
83130,0.0
|
| 131 |
+
83131,0.05263157894736842
|
| 132 |
+
83132,0.05263157894736842
|
| 133 |
+
83133,0.5263157894736842
|
| 134 |
+
83134,0.3684210526315789
|
| 135 |
+
83135,0.0
|
| 136 |
+
83136,0.15789473684210525
|
| 137 |
+
83137,1.631578947368421
|
| 138 |
+
83138,0.2631578947368421
|
| 139 |
+
83139,0.3684210526315789
|
| 140 |
+
83140,0.8421052631578947
|
| 141 |
+
83141,0.21052631578947367
|
| 142 |
+
83142,0.15789473684210525
|
| 143 |
+
83143,0.5263157894736842
|
| 144 |
+
83144,0.15789473684210525
|
| 145 |
+
83145,0.47368421052631576
|
| 146 |
+
83146,0.3684210526315789
|
| 147 |
+
83147,0.10526315789473684
|
| 148 |
+
83148,0.42105263157894735
|
| 149 |
+
83149,0.3157894736842105
|
| 150 |
+
83150,0.3684210526315789
|
| 151 |
+
83151,0.0
|
| 152 |
+
83152,0.10526315789473684
|
| 153 |
+
83153,0.47368421052631576
|
| 154 |
+
83154,0.47368421052631576
|