Create a data.tree object to be used for aggregating or scaling hierarchical categorical or interval data

Creates a hierarchical data.tree object that defines how different levels of categorical or interval data are related to each other.

create_agg_tree(mapping, exists, col_type)

create_scale_tree(mapping, exists, col_type, collapse_missing = FALSE)

vis_tree(tree)

Arguments

mapping: [data.table()]
For 'categorical' variables, defines how different levels of the hierarchical variable relate to each other. For aggregating 'interval' variables, it is used to specify intervals to aggregate to, while when scaling the mapping is inferred from the available intervals in dt.
exists: [character()]
names of variables in the mapping that data exists for.
col_type: [character(1)]
The type of variable that is being aggregated or scaled over. Can be either 'categorical' or 'interval'.
collapse_missing: [logical(1)]
When scaling a categorical variable, whether to collapse missing intermediate levels in mapping. Default is 'False' and the function errors out due to missing data.
tree: [data.tree()] as returned by create_agg_tree() or create_scale_tree().

Value

create_agg_tree() and create_scale_tree() return a [data.tree()] with attributes for whether each node has data available ('exists') and whether aggregation to or scaling of each node is possible ('agg_possible' or 'scale_possible'). For create_scale_tree(), also includes a field for whether children of each node can be scaled ('scale_children_possible').

vis_tree() uses networkD3::diagonalNetwork() to create 'D3' network graphs.

Details

Hierarchical tree data structures are used to represent different levels of the categorical or interval data. The r package data.tree is used to implement these data structures.

When vis_tree() is used to visualize a tree returned by create_agg_tree() then nodes with data directly provided are colored green, nodes where aggregation is possible are colored blue, and missing nodes without data directly provided and where aggregation is impossible because of the missing nodes are colored red.

When vis_tree() is used to visualize a tree returned by create_scale_tree() then nodes with data directly provided and that can be scaled are colored green, nodes with data directly provided but that can not be scaled are colored blue and nodes without data directly provided

Examples

# aggregation example where all present day locations exist except for Tehran
locations_present <- iran_mapping[!grepl("[0-9]+", child) &
                                   child != "Tehran", child]
agg_tree <- create_agg_tree(iran_mapping, exists = locations_present,
                            col_type = "categorical")
vis_tree(agg_tree)


# scaling example where all present day locations exist without collapsing
locations_present <- c(iran_mapping[!grepl("[0-9]+", child), child], "Iran")
scale_tree <- create_scale_tree(iran_mapping,
                                exists = locations_present,
                                col_type = "categorical")
vis_tree(scale_tree)


# scaling example where all present day locations exist and collapsing tree
scale_tree <- create_scale_tree(iran_mapping,
                                exists = locations_present,
                                col_type = "categorical",
                                collapse = TRUE)
vis_tree(scale_tree)