short-term-forecasting

Short-term forecasting of electricity generation, demand and prices using machine learning, by Nithiya Streethran (nmstreethran@gmail.com).

This is a work-in-progress. Please feel free to suggest improvements (see the code of conduct and contributing guidelines).

Table of contents

• Folders
• Documentation
• My specifications
• Funding
• Licenses and terms of use
• Credits

Folders

• data contains datasets (both input and output) and their terms of use (available externally on Dropbox)
• scripts contains all Python scripts
• jupyter-notebooks contains Python files with outputs in Jupyter notebook format
• docs contains the documentation and associated files
• images contains images and their license

Documentation

Documentation is written in the repository's GitHub Wiki. The docs folder contains the documentation and associated files in Markdown, PDF and HTML formats.

The file docs.sh contains shell commands to perform the document conversions, and also copies the files from the wiki (which is a separate repository) into this repository. It is executed in Bash using the following command:

bash docs.sh

The following programmes and packages are required to successfully execute the above script:

• Pandoc
• pandoc-citeproc
• pandoc-crossref
• Git Bash (if using Windows)
• LaTeX (or a TeX distribution, such as TeX Live)

The GitHub wiki has been included in this repository as a submodule using the following command:

git submodule add https://github.com/ENSYSTRA/short-term-forecasting.wiki.git wiki

The submodule allows the wiki to be cloned locally into the same directory as the main repository, which allows its inclusion in releases. Once changes to the wiki within the submodule are made (e.g., new markdown files, images), these changes are first committed and pushed to the wiki's branch, before committing and pushing to the main repository's branch.

The list of works cited can be found in Zotero, and in this repository as a BibTeX bibliography database (.bib) file.

The in-text citation keys used are automatically generated using Zotero with the Better BibTeX extension using the following format (changed via Edit > Preferences > Better BibTeX > Citation keys in the Zotero desktop application):

[auth5_1][>0][shortyear] | [title:substring=1,5][shortyear]

My specifications

The source code editor I use is VSCodium (fully open-source alternative to Visual Studio Code), with Git integration and the following useful extensions:

• Anaconda Extension Pack
• Python
• Visual Studio IntelliCode
• Code Spell Checker
• Markdown All in One
• GitLens — Git supercharged

My current computing specifications:

• Python version 3.6.8
• conda version 4.6.11
• git version 2.18.0.windows.1
• Processor: Intel(R) Core(TM) i5-7200U CPU @ 250 GHz 2.71 GHz
• RAM: 8 GB
• Default shell: Bash

Funding

This work is part of my research as Early-Stage Researcher (ESR) 9 of the ENSYSTRA (ENergy SYStems in TRAnsition) Innovative Training Network at the University of Stavanger. ENSYSTRA is funded by the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No: 765515.

   

      

            

Licenses and terms of use

Where sources have not been specified:

• Code license: GNU General Public License Version 3
• Documentation license: GNU Free Documentation License Version 1.3
• Image license: Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)
• Output data license: ODC Open Database License (ODbL)

Licenses and terms of the input data used can be found in their corresponding folders within the data folder on Dropbox. For more information, please see the documentation.

Credits

Contributing guidelines: adapted from the Open Science MOOC.

License badges: lukas-h/license-badges.md; made with Shields.io.

Markdown-formatted Creative Commons licenses: idleberg/Creative-Commons-Markdown.

IEEEurl.csl: Citation Style Language for Zotero, originally by Michael Berkowitz, Julian Onions, Rintze Zelle, Stephen Frank and Sebastian Karcher.

solarized.css: CSS for HTML documents, originally by Thomas Frössman, modified using some elements from pandoc.css by killercup. The Solarized color theme was originally developed by Ethan Schoonover.

pandoc.tex: LaTeX template for formatting PDFs generated using Pandoc and a LaTeX PDF engine, originally downloaded from Pandoc's demo page.