Devlog

A new way through the catalog: browse by language family

Eleven language-family hubs plus Bantu, Nilotic and Polynesian branch pages, and the open phenotype dataset rebuilt from 484 to 1,779 groups.

  • Catalog
  • Data
  • Research

On July 26 the catalog gained a new axis. The /cluster hubs group the peoples by language family: Niger-Congo, Indo-European, Austronesian, Sino-Tibetan, Afroasiatic, Nilo-Saharan, Turkic, Kra-Dai, Austroasiatic, Dravidian and Uralic. Three branch hubs go one level deeper, for the Bantu, Nilotic and Polynesian peoples.

Family hubs group their members by branch, and each branch heading links to its own hub where one exists. People pages now link up to their language family, and to their branch hub where they have one.

The same day the open phenotype dataset, EthnicErotic/phenotype-catalog, was rebuilt from 484 to 1,779 groups. It carries the full catalog, including 167 groups published for research use, and phenotype profiles for 1,777 groups.

Under the hood

  • Family membership is derived, not stored: it is computed at request time from each people's language classification path, written as family, branch, sub-branch and language, so the hubs follow the catalog as it grows.
  • The parser normalizes dash variants in family names such as Niger-Congo and Kra-Dai, and reads two path formats: a slash-separated path and a parenthetical, comma-separated one.
  • Family hubs match on the first path segment. Branch hubs match any segment, and the Polynesian hub also takes in peoples whose homeland region is Polynesia where the language path skips the branch level.
  • The dataset's ethnicities config carries normalized linguistic, geographic, cultural and phenotype metadata for every group, with a canonical link back to the site for each group that has a public page. A corpus_only column, new in schema v4, marked the 167 groups that the release carried as research rows.
  • Those 167 groups were selected for verifiable attestation in Wikipedia, ISO 639-3 language codes and named ethnography, with uncontacted and recently contacted peoples excluded. Each carries a phenotype profile that feeds the image pipeline, which rendered portraits for all 167 on July 26.
groups in the open dataset
1,779
language-family hubs
11
branch hubs
3

See it on the site

Sources and open data

Read the full log