koboloadeRlight-package: koboloadeRlight: Data crunching light for kobo forms
In ElliottMess/koboloadeRlight: Data crunching light for kobo forms

A package to conduct data discovery on Kobotoolbox forms and datasets.

Prerequisite
To use koboloadeRlight, you will need at least a dataframe that follows basic principales of Kobotoolbox exported files:

No group names in the column headers
Download the data with XLM values and headers
No special characters in column headers or choices, expect for select_multiple questions (see below). To do so, make sure that there is sepcial symbols in your XLSform before export.
Select_multiple questions should appear in the dataframe in at least two columns (this is the default option when exporting data from kobo):
- One generic column with the name of the question as header and all choices that were picked by respondents
- One column per choice in the question. The name of the column should be the name of the question, followed by a forward slash, and the name of the choice (e.g. ws_multi/pomp).

Walk through
koboloadeRlight can handle two main scenarii of data exploration

Data downloaded right from a Kobotoolbox server. Data and form are available.
Data from a kobotoolbox server modified. Typically cleaned data.

For the first scenario:

Import your data as a dataframe (with read_csv() or read_excel for instance).
Run kobo_dico() to build a dictionnary linking your form and your data.
Run a choice of kobo_crunching_all(), kobo_crunching() or other analysis functions.

If you don't have access to the updated form:

If you added questions to your dataframe (e.g. made additional calculations that you stored as columns), make sure to add them to your form.
Run kobo_to_xlsform() to generate a new form from the data
If you have a XLSform that genereated your data, you can use kobo_xlsform_relabel() to relabel your choices and questions according to the form.
Run kobo_dico() to build a dictionnary linking your form and your data.
Run a choice of kobo_crunching_all(), kobo_crunching() or other analysis functions.

This package is based on koboloadeR https://github.com/unhcr/koboloadeR. It provides a lighter alternative to koboloadeR, focusing on core functions.
The package is built for data exported from KoboToolbox. Full documentation for KoboToolbox is available here : https://www.kobotoolbox.org/

ElliottMess/koboloadeRlight documentation built on June 20, 2019, 2:24 p.m.