This repository is the reproducibility package for the article “Using mobile Devices as Scientific Measurement Instruments: Reliable Android Task Scheduling". In terms of software, the paper provides two main outputs. The first and most important is the NativeScript Task Dispatcher library, see https://github.com/GeoTecINIT/nativescript-task-dispatcher (10.5281/zenodo.4530103). The second is the analysis and results of one of the experiments reported in the article (Experiment 1: simple task scheduling), which is the focus of this repository.

Click the “Binder” button below to open an interactive editing environment with all required software installed on MyBinder.org. It uses the current version of the branch main in the repository, but you can also enter the Zenodo DOI (see above) in the MyBinder user interface to open a preserved release version.

You can start RStudio for the analysis via “New > RStudio”. You can navigate to the R Markdown notebook file (see list of files below) to inspect the code and reproduce (“Knit > Knit to HTML”) the table and figure as described in Reproduce analysis, except that local installation of required packages is not required.

Open the R Markdown file (ntd-experiment1.Rmd) with RStudio. Then select “Knit > Knit to HTML” to render the web page. If you have errors rendering the HTML page, try running each chunk to locate the problem. This analysis does not convert raw datasets into processed data for the analysis - please execute the chunk merge_datafiles manually if necessary.

The R Markdown file does not include code to install required packages. Run the code in the file install.R to install all dependencies.

Install Docker CE or a compatible tool for building an image based on a Dockerfile and running a container based on the image. The Dockerfile uses the Rocker image rocker/binder:3.6.3, providing R version 3.6.3 with a CRAN mirror timestamp of July 5th 2019.

Download the project files, open a command line in the root directory (where this file is), and run the commands as documented at the end of the Dockerfile file.

If you have repo2docker, you can also run repo2docker --image-name=rr-ntd-ieeetmc . The repo2docker option is the only way the original authors worked on the analysis to ensure the computing environment is properly managed.

• data-raw/*.csv. Each file corresponds to the raw data collected during 2 weeks per type of device and scheduling application.
• data/data_ieeetmc.{csv,rds}: Processed data ready for analysis that integrates all raw datasets.
• data/devices.csv: List of Android-based mobile devices used for experimentation.
• ntd-experiment1.Rmd: R Markdown document with the analysis and visualisation of the Experiment 1: simple task scheduling.
• figs/fig_boxplot.png: figure created by the analysis ntd-experiment1.Rmd and is part of the submitted article (Fig 4).
• Dockerfile: A recipe for the computational environment using Docker.
• install.R: R script file executed during creation of the Docker image to install required dependencies.

The documents in this repository are licensed under Creative Commons Attribution 4.0 International License.

All contained code is licensed under the Apache License 2.0.

The data used is licensed under Open Data Commons Attribution License.