One lane per axis. Horizontal position: share of a node's papers that study sources (left) or structure (right); papers with both targets pull toward the middle. Node area: number of papers; an edge joins two nodes carried by the same papers (curved within a lane). Two terms are in the same analysis when they belong to one method use of a paper, together with the networks that method draws on; otherwise they only share the paper. Terms that describe the whole study (contribution, tectonic setting, focal event, spatial scope, time design, data origin, source–structure link) are not tied to one analysis: their edges are dashed in the share view and left out of the same-analysis view. In the same-analysis view, observed / expected counts analyses instead of papers. Hover a node to see its partners. Scroll to zoom in: nodes and labels grow up to about 1.5×, then keep that size while nodes spread apart and more labels appear; drag to pan.
Observed / expected = N·n(a,b) / (n(a)·n(b)) within the current filters, with N papers in view; above 1 means the pair co-occurs more often than chance. In the same analysis: a cell counts the papers that use both terms in one analysis (one method use with the networks it draws on, or one network), and observed / expected counts analyses instead of papers, M·m(a,b) / (m(a)·m(b)) with M analyses in view. Click a cell for its papers. Drag a row or column name to move it.