Implementation:Online ml River Datasets Restaurants

Knowledge Sources	Online_ml_River
Domains	Online_Learning, Datasets, Regression, Time_Series
Last Updated	2026-02-08 16:00 GMT

Overview

Concrete dataset for regression provided by the River library.

Description

Data from the Kaggle Recruit Restaurants challenge. The goal is to predict the number of visitors in each of 829 Japanese restaurants over a period of roughly 16 weeks. The data is ordered by date and then by restaurant ID.

This dataset contains 252,108 samples with 7 features for regression tasks.

Usage

This dataset is useful for:

• Time series forecasting with multiple entities
• Restaurant visitor prediction
• Demand forecasting in hospitality industry
• Multi-entity time series problems

Code Reference

Source Location

• Repository: Online_ml_River
• File: river/datasets/restaurants.py

Signature

class Restaurants(base.RemoteDataset):
    def __init__(self):
        super().__init__(
            n_samples=252_108,
            n_features=7,
            task=base.REG,
            url="https://maxhalford.github.io/files/datasets/kaggle_recruit_restaurants.zip",
            size=28_881_242,
            filename="kaggle_recruit_restaurants.csv",
        )

    def _iter(self):
        return stream.iter_csv(
            self.path,
            target="visitors",
            converters={
                "latitude": float,
                "longitude": float,
                "visitors": int,
                "is_holiday": ast.literal_eval,
            },
            parse_dates={"date": "%Y-%m-%d"},
        )

Import

from river import datasets
dataset = datasets.Restaurants()

I/O Contract

Inputs

Name	Type	Required	Description
(none)	—	—	No parameters needed

Outputs

Name	Type	Description
iter()	tuple(dict, int)	Yields (features_dict, target) pairs where target is visitor count

Dataset Properties

Property	Value
Number of samples	252,108
Number of features	7
Task	Regression
Format	CSV (compressed)
Size	28,881,242 bytes
Number of restaurants	829
Time period	~16 weeks

Features

The dataset includes the following features:

• date: Date of the observation (datetime)
• latitude: Restaurant latitude (float)
• longitude: Restaurant longitude (float)
• is_holiday: Whether the date is a holiday (boolean)
• Restaurant identification features
• visitors: Number of visitors (target variable, integer)

Usage Examples

from river import datasets

dataset = datasets.Restaurants()
for x, y in dataset:
    print(x, y)
    break

References

• Recruit Restaurant Visitor Forecasting

Related Pages

• Environment:Online_ml_River_Python_Runtime_Environment