lprisan / ieeetlt-orchestration-load

Preprocessing, analysis and visualization scripts for the IEEE TLT journal paper, "Orchestration Load Indicators and Patterns" by Prieto, Sharma & Dillenbourg.

ieeetlt-orchestration-load

Preprocessing, analysis and visualization scripts for the IEEE TLT journal paper, "Orchestration Load Indicators and Patterns" by Prieto, Sharma & Dillenbourg.

Basically, this is a set of R markdown reports and scripts to help reproduce the download of raw data (from Zenodo -- Study 1, Study 2, Study 3, Study 4), data pre-processing, analysis and visualizations, many of which are directly used in the paper. All these processes are tied together by reports for each case, and another cross-case analysis report in R Markdown, to be compiled using the knitr package.

Requirements and Installation

In order to reproduce the paper's data analysis and generate the corresponding report, you need at least the following:

• An internet connection (to download the data)
• The curl tool (to download the data)
• R
• RStudio (optional, but very convenient)
• The knitr package

Usage

Basically, open the desired .Rmd file in RStudio, and click on Knit HTML. An HTML version of the report should be generated (this may take some time, though!).

Project structure

The project is organized in a folder structure, inspired by ProjectTemplate:

• data/ : Contains the scripts to download, uncompress and load the raw data
• lib/ : Contains some useful R scripts used throughout the report, such as the functions to calculate rolling windows, extract extreme load episodes to be video coded, etc.