Skip to contents

Topic segmentation sunburst chart

Source: R/rdb.gen.R
plot_topic_segmentation.Rd

Creates a Plotly sunburst chart that visualizes the hierarchical segmentation of referendum topic occurences.

Usage

plot_topic_segmentation(
  data,
  method = c("per_rfrnd", "per_topic_lineage", "naive")
)

Arguments

data

RDB referendum data as returned by rfrnds(). A data frame that at minimum contains the columns topics_tier_1, topics_tier_2 and topics_tier_3.

method

Applied method to count the number of topic occurences. One of

  • "per_rfrnd": All referendums have the same weight. For a referendum with n different topics of the same tier, every topic is counted 1/n.

  • "per_topic_lineage": All topic lineages have the same weight. For a referendum with n different topics of the same tier, every topic is fully counted, meaning that e.g. a referendum with three different tier-3 topics has a tripled impact on the result compared to a referendum that only has a single one. Noticeably faster than "per_rfrnd".

  • "naive": Naive procedure which doesn't properly reflect topic proportions on tier 2 and 3. Based on the (wrong) assumptions that a) all referendums have the same number of topic lineages assigned and b) topics are not deduplicated per tier. By far the fastest method, though.

Value

A plotly object.

Details

A topic lineage is the hierarchical compound of a topic_tier_1 and optionally a grandchild topic_tier_3 and/or a child topic_tier_2.

Note that topics can be assigned on any tier to referendums (i.e. in one case, a topic_tier_1 plus a child topic_tier_2 is assigned, and in another case only a topic_tier_1 without any further child topic).

Furthermore, it should be noted that not every topic_tier_2 has potential child topic_tier_3s. See the full topic hierarchy for details.

See also

Other visualization functions: ggplot_streamgraph(), plot_rfrnd_share_per_period(), plot_topic_share_per_period()

Examples

# count each referendum equally
rdb::rfrnds(quiet = TRUE) |>
  rdb::plot_topic_segmentation(method = "per_rfrnd")



# count each topic lineage equally
rdb::rfrnds(quiet = TRUE) |>
  rdb::plot_topic_segmentation(method = "per_topic_lineage")



# naive count (way faster, but with misleading proportions on tier 2 and 3)
rdb::rfrnds(quiet = TRUE) |>
  rdb::plot_topic_segmentation(method = "naive")