---
license: mit
task_categories:
- image-classification
language:
- de
tags:
- food-waste
- nutrition
- image-data
dataset_info:
  features:
  - name: bonid
    dtype: int64
  - name: image
    dtype: image
  - name: Bon_ID
    sequence: int64
  - name: Artikelnummer
    sequence: int64
  - name: Artikel
    sequence: string
  - name: Stückartikel
    sequence: string
  - name: Anzahl_Kellen
    sequence: int64
  - name: Gewicht_Kelle
    sequence: int64
  - name: Gewicht_Teller
    sequence: int64
  - name: kcal_Teller
    sequence: string
  - name: kj_Teller
    sequence: string
  - name: Fett_Teller
    sequence: string
  - name: ges_Fettsäuren_Teller
    sequence: string
  - name: Kohlenhydrate_Teller
    sequence: string
  - name: Zucker_Teller
    sequence: string
  - name: Eiweiß_Teller
    sequence: string
  - name: Salz_Teller
    sequence: string
  - name: Menge_Rückläufer
    sequence: float32
  - name: Prozent_Rückläufer
    sequence: float32
  - name: Gericht
    dtype: string
  - name: Portionsgröße
    dtype: string
  - name: Gewicht_vorher
    dtype: int64
  - name: kcal_vorher
    dtype: string
  - name: kj_vorher
    dtype: string
  - name: Fett_vorher
    dtype: string
  - name: ges_Fettsäuren_vorher
    dtype: string
  - name: Kohlenhydrate_vorher
    dtype: string
  - name: Zucker_vorher
    dtype: string
  - name: Eiweiß_vorher
    dtype: string
  - name: Salz_vorher
    dtype: string
  - name: Gewicht_nachher
    dtype: int64
  - name: kcal_nachher
    dtype: string
  - name: kj_nachher
    dtype: string
  - name: Fett_nachher
    dtype: string
  - name: ges_fettsäuren_nachher
    dtype: string
  - name: Kohlenhydrate_nachher
    dtype: string
  - name: Zucker_nachher
    dtype: string
  - name: Eiweiß_nachher
    dtype: string
  - name: Salz_nachher
    dtype: string
  splits:
  - name: train
    num_bytes: 51270168.0
    num_examples: 215
  - name: test
    num_bytes: 77496685.0
    num_examples: 160
  download_size: 51069023
  dataset_size: 51270168.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---

# Food Waste Dataset

This dataset contains detailed information about food waste, including images and nutritional information.

## Dataset Description

The dataset combines image data with detailed nutritional information for each meal and its ingredients.

### Features

For each entry:
- `bonid`: Unique identifier for each meal
- `image`: Image of the meal
- Lists per meal (multiple ingredients):
  - `Artikelnummer`: Article numbers
  - `Artikel`: Ingredient names
  - `Stückartikel`: Piece article information
  - `Anzahl_Kellen`: Number of portions
  - `Gewicht_Kelle`: Weight per portion
  - `Gewicht_Teller`: Weight per plate
  - `kcal_Teller`, `kj_Teller`: Caloric information
  - `Fett_Teller`, `ges_Fettsäuren_Teller`: Fat content
  - `Kohlenhydrate_Teller`, `Zucker_Teller`: Carbohydrate content
  - `Eiweiß_Teller`: Protein content
  - `Salz_Teller`: Salt content
  - `Menge_Rückläufer`, `Prozent_Rückläufer`: Return quantities

Per image measurements:
- Before consumption:
  - `Gewicht_vorher`: Initial weight
  - `kcal_vorher`, `kj_vorher`: Initial calories
  - `Fett_vorher`, `ges_Fettsäuren_vorher`: Initial fat content
  - `Kohlenhydrate_vorher`, `Zucker_vorher`: Initial carbohydrates
  - `Eiweiß_vorher`: Initial protein
  - `Salz_vorher`: Initial salt
- After consumption:
  - `Gewicht_nachher`: Remaining weight
  - `kcal_nachher`, `kj_nachher`: Remaining calories
  - `Fett_nachher`, `ges_fettsäuren_nachher`: Remaining fat
  - `Kohlenhydrate_nachher`, `Zucker_nachher`: Remaining carbohydrates
  - `Eiweiß_nachher`: Remaining protein
  - `Salz_nachher`: Remaining salt

## Usage

```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("AI-ServicesBB/food-waste-dataset")

# Access an example
example = dataset['train'][0]

# Get image and nutritional data
image = example['image']
ingredients = example['Artikel']
weights = example['Gewicht_Teller']
calories = example['kcal_Teller']

# Get before/after measurements
weight_before = example['Gewicht_vorher']
weight_after = example['Gewicht_nachher']
```

## Dataset Creation

The dataset combines two main sources:
1. Ingredient-specific measurements for each meal
2. Before/after measurements with images for complete meals

All numeric values are standardized to use decimal points (converted from German number format).

## Additional Information

- Total number of entries: {len(dataset)}
- Image format: JPEG
- Measurements are in grams, calories, and standard nutritional units