Takes a coin and outputs an analysis of indicators. The objective is to look for any "statistically-problematic" indicators, based on:
f_analyse_indicators(coin)
The coin
coin Updated coin with analysis tables
data availability
high proportion of repeated data values
outliers as defined by skew and kurtosis
collinearity within first aggregation level (category)
negative correlation within first aggregation level (category)
The output can be viewed in R or as an interactive data frame. Uses Spearman rank correlation to deal with skewed distributions. Operates on the Raw data set.
The output is a coin with the results attached. This is so that on export, the results will also be exported easily.