The Missing Path

Research, Design & Development

2020 -

Marie designed and developed The Missing Path, a visualisation tool to support data producers in identifying and analysing incompleteness in Knowledge Graphs.
It computes clusters with similar incomplete profiles on a map and lets them inspect and contextualize their statistical summaries. They gain insights on incompleteness and find strategies to identify coherent subsets to be fixed.

The map on the left shows the 4567 Comics entities in Wikidata. The clusters appearing represent groups of entities that share the same missing paths. If a collection were 100% complete, there would be only one large cluster. The user has selected a small cluster of 20 entities on the left of the map; it is colored in dark pink. On the left column, the histogram of paths completeness for the full collection can be compared with the histogram for the selected subset on the right. Each row represe
Four maps representing four wikidata collections. The number of clusters, their size and distribution provide a visual footprint of the shape of a collection, relative to the set of paths selected to produce the map (highlighted in pink on the right side of each thumbnail).