b-fatma/bike-sharing-federated

---
dataset_info:
features:
- name: yr
dtype: int64
- name: workingday
dtype: int64
- name: atemp
dtype: float64
- name: hum
dtype: float64
- name: windspeed
dtype: float64
- name: hr_sin
dtype: float64
- name: hr_cos
dtype: float64
- name: season_2
dtype: int64
- name: season_3
dtype: int64
- name: season_4
dtype: int64
- name: weathersit_2
dtype: int64
- name: weathersit_3
dtype: int64
- name: cnt_log
dtype: float64
splits:
- name: train
num_bytes: 1445600
num_examples: 13900
- name: test
num_bytes: 361504
num_examples: 3476
download_size: 127096
dataset_size: 1807104
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
---
---

## license: cc-by-4.0

# Bike Sharing Dataset (Processed for Federated Learning)

## Description

This dataset is a processed version of the Bike Sharing dataset from the UCI Machine Learning Repository.
It has been cleaned, encoded, and prepared for machine learning and federated learning experiments.

The dataset contains hourly bike rental data along with weather and seasonal information.
The target variable represents the total number of bike rentals.

## Source Dataset

Fanaee-T, H. (2013). *Bike Sharing Dataset*. UCI Machine Learning Repository.
https://doi.org/10.24432/C5W894
https://archive.ics.uci.edu/dataset/275/bike+sharing+dataset

## Preprocessing Steps

The original dataset was processed as follows:

* Removed date column (`dteday`)
* One-hot encoded categorical features:

* `season`
* `weathersit`
* Cyclical encoding applied to:

* `hr` → `hr_sin`, `hr_cos`
* Target variable transformed:

* `cnt_log = log(1 + cnt)`
* Train/test split created
* Dataset formatted for PyTorch and federated learning experiments

## Features

The dataset includes the following features:

* yr
* workingday
* atemp
* hum
* windspeed
* hr_sin
* hr_cos
* season_2
* season_3
* season_4
* weathersit_2
* weathersit_3
* weathersit_4

## Target

* **cnt_log**: Log-transformed total number of bike rentals

## Dataset Splits

* **train**: Training data
* **test**: Test data

## Intended Use

This dataset is intended for:

* Regression tasks
* Federated learning experiments
* PyTorch model training
* Feature engineering and ML pipelines

## License

This dataset is distributed under the **CC BY 4.0** license, following the license of the original dataset.