# Finding other Uruguay data Cows are a beginning. Uruguay's national open-data catalogue, catalogodatos.gub.uy, lists 2,718 datasets from 70 publishers (harvested on 14 October 2026), and the graph `uruguay-catalogue` holds every one of them: title, description, publisher, themes, keywords, licence, dates, and each of their 14,426 files with its download link. Of the 2,718, 1,875 come from the Ministerio de Desarrollo Social (MIDES), most of them one indicator each; the Intendencia de Montevideo has 159, ANCAP 60, the national post office 43. Somebody, somewhere, has published the readings of a meteorological buoy, the price of fuel at the refinery gate, and the number of women mayors in each department. They are all here, waiting. ## By meaning The graph's documents are cards: one per publisher (MIDES needs thirteen), one paragraph per dataset. Each paragraph has the Spanish title, an English gloss of it, the publisher's description, themes in both languages, the file formats, the last modified date, and the dataset's IRI. Semantic search runs over those paragraphs, so a question in either language finds them: ``` semantic_search graphId: uruguay-catalogue query: "electricity consumption by department" semantic_search graphId: uruguay-catalogue query: "consumo de energía" semantic_search graphId: uruguay-catalogue query: "school enrolment" ``` Each hit's `content` ends with `Dataset : >`. That IRI is the dataset in the DCAT graph. Search finds what exists, which is not always what was asked for. There is, for instance, no series of electricity consumption by department on the catalogue: the national energy balance (MIEM's *Balance Energético Nacional*, BEN) gives consumption by sector and by source for the whole country, and MIDES gives the share of households without electricity by department. A good answer says so, and offers both. The English glosses are Sophia's translations of the titles, not the publishers'. They exist because the semantic model is better at English than at Spanish; the Spanish is always beside them, and it is the Spanish that is official. ## By name or publisher `search_documents` matches card titles (publishers), and SPARQL matches anything else. The catalogue is in the named graph `urn:sophia:garden:uruguay-catalogue:dcat` (always wrap patterns in `GRAPH`): ```sparql PREFIX dcat: PREFIX dct: SELECT ?dataset ?title WHERE { GRAPH { ?dataset a dcat:Dataset ; dct:title ?title ; dct:publisher . } } ORDER BY ?title ``` ## From a dataset to its files ```sparql PREFIX dcat: PREFIX dct: SELECT ?file ?format ?url ?modified WHERE { GRAPH { dcat:distribution ?d . ?d dct:title ?file ; dct:format ?format ; dcat:accessURL ?url . OPTIONAL { ?d dct:modified ?modified } } } ``` `dcat:downloadURL` is set when the file is an upload or a file format (CSV, XLSX, JSON, ZIP, SHP and the like); a link to an API or a web page has only `dcat:accessURL`. The files themselves are not in the graph: fetch them from the publisher's URL. ## What else is recorded - **Themes:** the catalogue's 22 groups (`dcat:theme`, SKOS concepts with Spanish and English labels): Statistics, Transparency, Education, Health, Agriculture and so on. - **Licences:** 2,664 datasets carry the Licencia de Datos Abiertos – Uruguay; the rest CC0, CC BY, CC BY-SA, one CC BY-NC, and 24 no licence stated (`dct:license`). - **Frequency:** the catalogue's update frequency in days (`uycat:updateFrequencyDays`), and the EU frequency term where the number has an obvious one. - **Provenance:** the named graph `urn:sophia:garden:uruguay-catalogue:provenance` records the harvest (date, requests, the sha256 of every raw answer) and the attribution the licence asks for. ## Citing what you find The licence asks anyone who uses the data or the metadata to name the provider, the "Licencia de Datos Abiertos – Uruguay", and the dataset, and to say what they changed. For the catalogue itself: *Fuente: Catálogo Nacional de Datos Abiertos (catalogodatos.gub.uy), metadatos de cada organismo proveedor, Licencia de Datos Abiertos – Uruguay; cosechados el 14/10/2026 y convertidos a DCAT por Sophia Labs, con glosas en inglés no oficiales.* For a dataset: its publisher, its title and its catalogue address. The catalogue changes; this graph does not, until the next harvest. The date on the cards is the date of the world they describe. --- *This guide is by Sophia Labs, under CC BY 4.0. The data it describes is MGAP's (and, for the catalogue, each publisher's), under the Licencia de Datos Abiertos – Uruguay.* The cow-tools guide, The guide: https://cow-tools.sophia-labs.com/guide/finding-other-uruguay-data/ (this page for people) · the whole guide: https://cow-tools.sophia-labs.com/guide/index.md · in Sophia: the graph `cow-tools-guide`, document `finding-other-uruguay-data` (MCP `read_document` at https://cow-tools.sophia-labs.com/mcp). CC BY 4.0, credit "cow-tools, Sophia Labs".