Wrocław Open Dataset - WRM usage predictor

ML project to predict usage of WRM

Getting Started

Data input file (compressed with 7z because standard zip had over 130MB which is 1/7 of the original file):

Download data/model_input.7z

Prerequisites

Packages:

tensorflow>=2.1
scikit-learn
seaborn>=0.10.0
pandas>=1.0.1
notebook>=6.0.3
matplotlib>=3.2.0
jupyter-core>=4.6.3
xlrd>=1.2.0

Installing

Best way to install dependencies and avoid unnecessary problems is to setup Anaconda env and run following command inside the environment

pip install -r requirements.txt

after that you should be able to run:

jupyter notebook

to open one of the notebooks.

Give an example

And coding style tests

Explain what these tests test and why

Give an example

Data Processing

• Weather Data Source
• Usage Source

For weather data description check Files description This Link

For each day we're extracting data in format:

[
    'date',
    'time',
    'totalSnow_cm',
    'sunrise',
    'sunset',
    'tempC',
    'FeelsLikeC',
    'HeatIndexC',
    'windspeedKmph',
    'weatherCode',
    'precipMM',
    'humidity',
    'visibility',
    'pressure',
    'cloudcover'
]

Most of the data comes directly from the source except:

• time - source data converted from string 1300 to number 13*60
• sunrise - date converted from HH:mm AM to minutes, started from midnight
• sunset - date converted from HH:mm PM to minutes, started from midnight

To run data processing:

Process weather data

python weather_parser.py

Generate WRM data per year and list of places in data/bike_data directory

python data_parser.py

Generated dataset has format of:

    'bike_number',
    'start_time',
    'end_time',
    'rental_place',
    'return_place',
    'year',
    'week_day',
    'totalSnow_cm',
    'sunrise',
    'sunset',
    'tempC',
    'FeelsLikeC',
    'HeatIndexC',
    'windspeedKmph',
    'weatherCode',
    'precipMM',
    'humidity',
    'visibility',
    'pressure',
    'cloudcover'

WARNING!!! It takes a while to run.

Deployment

Add additional notes about how to deploy this on a live system

Built With

• Tensorflow - An end-to-end open source machine learning platform for everyone
• Pandas - Open source data analysis and manipulation tool
• WOD - Wroclaw Open Dataset

Versioning

We use SemVer for versioning. For the versions available, see the https://github.com/burnpiro/wod-usage-predictor/tags.

Authors

See also the list of contributors who participated in this project.

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Acknowledgments

• Hat tip to anyone whose code was used
• Inspiration
• etc