Starting kit on the bike counters dataset

Read the instruction from the Kaggle challenge.

Download the data

Download the data from Kaggle and put the files into the data folder.

Note that your goal is to train your model on train.parquet (and eventual external datasets) and then make predictions on final_test.parquet.

Install the local environment

To run the notebook locally you will need the dependencies listed in requirements.txt.

It is recommended to create a new virtual environement for this project. For instance, with conda,

conda create -n bikes-count python=3.10
conda activate bikes-count

You can install the dependencies with the following command-line:

pip install -r requirements.txt -U

The starter notebook

Get started on this challenge with the bike_counters_starting_kit.ipynb notebook. This notebook is just a helper to show you different methods and ideas useful for your exploratory notebooks and your submission script.

Launch the notebook using:

jupyter lab bike_counters_starting_kit.ipynb

Submissions

Upload your script file .py to Kaggle using the Kaggle interface directly. The platform will then execute your code to generate your submission csv file, and compute your score.

Note that your submission .csv file must have the columns "Id" and "bike_log_count", and be of the same length as final_test.parquet.