{kobocruncher}

An organized workflow generating 'Rmd' files from an extended 'xlsform' questionnaire structure to facilitate survey data crunching.

{kobocruncher} support organised data analysis workflow, to conduct data discovery and analysis for data collected through KoboToolbox, ODK, ONA or any xlsform compliant data collection platform.

You can use this tool through a dedicated shiny App: https://rstudio.unhcr.org/kobocruncher/

This R package first builds on the capacity of UNHCR Kobo server but it can also be used from any structured dataset. It comes as a companion tool to the Integrated Framework for Household Survey.

Tutorial

{kobocruncher} aims at helping humanitarian data analysts to focus in data interpretation by saving the time needed to quickly generate the graphs and charts required to discover insights from a dataset. It also ensure analysis reproducibility through a separation of the analysis configuration and the analysis process. The package allows to account for sample weights and hierarchical dataset structure (both capacities that are not available through the default reporting engine or the excel-analyzer).

Presentations / tutorials for a one day training workshop are accessible here:

• 00-Intro
• 01-First Report
• 02-Relabeling
• 03-Grouping Questions
• 04-Setting Crosstabulation
• 05-Searching Asssociation
• 06-Indicator Calculation
• 07-Weighting
• 08-Publishing

Install

Beside the shinyApp, you can install the package on Rstudio.

If you are on Windows, you will first need to install Rtools on the top of R and Rstudio in order to install the package locally.

install.packages("pak")
pak::pkg_install("edouard-legoupil/kobocruncher")  

Configure authentication token

The {riddle} package is used to ensure integration with RIDL - UNHCR Internal Data Repository. It requires you to add your API token and store it for further use. The easiest way to do that is to store your API token in your .Renviron file which is automatically read by R on startup.

You can retrieve your API TOKEN in your user page.

To use the package, you’ll need to store your RIDL API token in the RIDL_API_TOKEN environment variable. The easiest way to do that is by calling usethis::edit_r_environ() and adding the line RIDL_API_TOKEN=xxxxx to the file before saving and restarting your R session.

Contributing

Contributions to the packages are welcome. Please read first the contribution guidelines, follow the code of conduct and use the issue template.

References

• JIPS Essential Toolkit, "How do we process and prepare data for analysis": https://jet.jips.org/wp-content/uploads/Guidance-Data-Processing-Phase5-JET.pdf

The package falls into the category of Automatic Data Exploration package (see an extensive review here) and mostly act as a wrapper for different specialized packages from the tidyverse .

Compare to those packages, {kobocruncher} provides:

• An integration of all the data processing stages, not only visualization

• An approach where most of the exploration is documented in the original xlsform used to collect the data as UNHCR standard for Survey Data Collection is based on Kobotoolbox. This allows for reproducible analysis

• Multiple notebook templates with standard charts designed to be interpretable by a non-technical audience

• Relabeling functions and Questions grouping ability

• Full integration with UNHCR Brand Style