synthdata: Synthetic code-transition network (Trees of Thought stand-in)

In Dynet: Tidy Temporal Network Analysis

Synthetic code-transition network (Trees of Thought stand-in)

Description

A synthetic temporal network of how one interaction code follows another in asynchronous discussion. Vertices are the ten codes a message can carry, so a vertex is a category, never a person, and a tie runs from the code of a message to the code of the message it replies to.

Usage

synthdata

Format

A data.frame with 101 rows and 5 columns:

from

Character. Code of the replying message.

to

Character. Code of the message replied to.

start

Numeric. When the spell opens, 0 to 2.10.

end

Numeric. When it closes, 0.03 to 5.80.

weight

Numeric, whole-valued. Message pairs the spell represents, 1 to 5881.

The ten codes are Acceptance, Argument, Composing, Disagreement, Evaluation, Group_regulation, Question, Sharing, Socioemotional and T.Regulation; 14 rows are self-loops.

Details

This is a stand-in for the network analysed in the Trees of Thought study, whose own data is not redistributable. It was built by drawing a random 70% subset of that study's edge spells and resampling that subset with replacement to 90% of the original spell count, so it omits about a third of the real spells and repeats others. No row of it should be read as a finding about the study, and figures computed from it will not match the published ones. It exists so the analysis can be run and taught end to end.

Spells are weighted by how many message pairs they represent, overlap in time, and include self-loops, because a code following itself is a real and common transition. Build with loops = TRUE to keep them; the weight column is picked up as the tie weight.

Source

Synthesised from the Trees of Thought code-transition network by the resampling described above; see data-raw/synthdata.R.

Examples

dynet(synthdata, directed = TRUE, loops = TRUE, weight = "weight")

Dynet documentation built on Oct. 7, 2026, 5:08 p.m.