Data Analysis Plan within your xlsfrom

Edouard Legoupil

2021-02-04

Xlsform is a convenient format to build advance form using any spreadsheet software such as Libreoffice or MsExcel.

In order to build an an analysis plan within the form, the columns described in the tables below needs to be added. Note that if the column are not present, the script will create dummy ones. It’s always possible to add your analysis plan to an existing form and relaunch kobo_dico in order to regenerate the correct analysis plan.

Note that for charting purpose, it’s recommanded that labels for questions & choices should not exceed 70 characters. It’s possible again to re-edit directly your xlsform and regenerate a new dico.

In the survey worksheet:

Column	Description
chapter	used to breakfdown the final report
disaggregation	used to flag variables used to facet dataset
correlate	used to flag variables used for statistical test of independence (for categorical variable) or correlation for numeric variable
variable	used to flag ordinal variables so that graphs are not ordered per frequency.
anonymise	used to generate an anonymised datset in line the anonymisation plan within the xlsform
structuralequation	used to tag variables to the standard structural equation model: risk, coping, vulnerability
clean	used to flag external csv file to be used for the cleaning of a specific variable
cluster	used to flag variables used for statistical clustering
predict	used to flag variables to be predicted based on a joined registration dataset
mappoint	used to flag variables to be mapped as point
mappoly	used to flag variables to be mapped as polygon

In the choices worksheet:

Column	Description
order	used to define order for ordinal variables
weight	used to define weight for each answers in case it’s used for some specific indicator calculation
recategorise	used to recategorise quickly choices for a question

In a separate indicator worksheet:

The idea is to map calculation necessary to create complex indicators from the variables defined in the survey worksheet. This will document and automate the generation of indicators (i.e. feature enginerring). This worksheet will allow to generate an additional elements to the data dictionnnary – i.e. dico dataframe.

Below are the minimum elements/column to document:

Column	Description
type	type of indicator select_one or integer or numeric
name	short name for the indicator
label	label for the indicator
chapter	chapter to include the indicator in
frame	frame to use to append the idnicator to
calculation	used to reference the calculation. This will be a precise R formula

define wether the indicator is Measurement: variable used to quantify other indicators, Disaggregation: variable that describes certain groups, Predictor: Indicator that describes the cause of a situation, Outcome: Indicator that describes the consequence of a situation or Judgment: indicator that translates a subjective assessment method to be used for the indicator: Percentage, Sum, Max/Min, Average, Score, Denominator, Numerator, Numerator.external (i.e. linked to an external value)