## Craft Brewery Panel Scoring Dataset

This dataset contains physicochemical measurements from a simulated craft brewery quality control program.  
Each row represents a beer batch evaluated by an expert sensory panel on a 0–5 quality scale.  
The data is intentionally **noisy, imbalanced, and partially missing** to mimic real-world production data.

***

### Files

- **train.csv** – training set with features and target `panel_score`  
- **test.csv** – test set with the same features (no `panel_score`)  
- **sample_submission.csv** – example submission file (`id`, `panel_score`)  

***

### Columns

All files share the same feature columns.  
Only `train.csv` contains the target `panel_score`.

**Identifiers**

- `id` – unique batch ID (e.g., `BR0001`, `BR0002`, …)

**Physicochemical features**

These are continuous numeric features derived from lab measurements:

- `base_malt_acidity` – fixed acidity from malt (g/L)  
- `fermentation_volatiles` – volatile acidity from fermentation by‑products (g/L)  
- `citric_additive` – citric acid additive level (g/L)  
- `post_ferment_sugar` – residual sugar after fermentation (g/L)  
- `salt_content` – salt (chloride) concentration (g/L)  
- `free_so2_protectant` – free SO₂ used as antioxidant (mg/L)  
- `total_so2_stabilizer` – total SO₂ used as preservative (mg/L)  
- `final_gravity` – final density of the beer (g/cm³)  
- `final_ph` – final pH of the beer  
- `sulfate_ion_level` – sulfate ion concentration (g/L)  
- `abv_alcohol_content` – alcohol by volume (% ABV)  
- `brew_temp_c` – fermentation temperature in °C  


**Target**

- `panel_score` – expert panel quality score in the range **0–5** (integer, ordinal)   
  - 1: poor  
  - 2: below average  
  - 3: average  
  - 4: good  
  - 5: excellent