"GRAND" Network Facts Label

Summarize your network.

A "Network Facts" label offers a fun way to summarize details of network data for a poster or presentation. However, it may not include all the details recommended by the Guidelines for Reporting About Network Data (GRAND). For more information about GRAND, visit https://www.grand-statement.org.

Examples

Small unipartite network
Large bipartite network
Longitudinal panel

Network

Network name

Data source?Empirical data may be collected using a variety of methods. Synthetic data is created by a generating model.

Received ethics approval

Location?The location that the data are intended to represent. Examples of possible locations include institutional locations (e.g., a company, a school), geographic locations (e.g., a country, a city), and digital locations (e.g., a social media platform).

Time?The time(s) that the data are intended to represent. This is not necessarily the same as when the data were collected or compiled.

Multiple?A single network dataset can contain multiple temporal waves of the same network (i.e., a panel), a network from each of multiple sites (e.g., communication networks in 10 companies), or a network from each of multiple egos (e.g., personal social networks of 10 people).

Availability?A DOI or URL to a repository containing these data.

Doesn't look like a DOI or URL — check for typos.

Citation?A DOI or URL to a manuscript describing these data.

Doesn't look like a DOI or URL — check for typos.

This network contains

Self-loops?A self-loop is an edge that connects a node to itself.

Multi-edges?When a network contains multi-edges, two nodes can be connected by more than one edge of the same type.

Missing data?Missing data can occur when data was not collected for nodes that should be included in the network, or when the presence/absence of an edge between a pair of nodes was not assessed.

Degree constraints?Degree constraints can arise by data collection (e.g., name up to 5 friends) or model specification (e.g., in the Barabasi & Albert (1999) model, m defines nodes’ minimum degree).

Hyper-edges?A single hyper-edge can simultaneously connect more than two nodes.

Nodes

Edges

Transformations

Symmetrized?Directed edges were transformed into undirected edges. Examples of symmetrizing transformations include confirmed edges (i.e., transforming only reciprocated edges i ⇄ j into undirected edges i−j) and unconfirmed edges (i.e., transforming any directed edges i →j and i ←j into undirected edges i−j).

Dichotomized?Weighted edges were transformed into unweighted edges. Examples of dichotomizing transformations include thresholds (i.e., count all edges with weight > X as present) and backbone models (i.e., count edges with statistically significant weights as present).

Projected?The network was generated from other data such that edges represent similarities, co-occurences, or statistical associations. Examples of such transformations include two-mode projection (e.g., two people are connected if they attended the same event) and (weighted) correlation networks (e.g., two variables are connected to the extent that they are correlated in a sample).

Excluded?Nodes or edges present in raw data were excluded from the analytic network. Nodes might be excluded for a variety of both structural (e.g., they are isolates with degree = 0, they are members of a small component) and non-structural (e.g., uncertainty, implausible responses) reasons.

Aggregated?Nodes or edges that were separate in raw data were aggregated in the analytic network. Individual nodes are sometimes aggregated into “supernodes” when they are structurally equivalent or belong to the same community. In multiplex networks, edges connecting the same dyad (e.g., i and j are connected by friendship and kinship) are sometimes aggregated into a single edge. In longitudinal data, edges occurring at specific times are sometimes aggregated into discrete waves.

Imputed?Nodes or edges that were unmeasured in raw data, and their presence or absence was imputed or inferred in the analytic network.