Dashboard: published water datasets
Every dataset in the Water group on data.ca.gov and the Water group on data.cnra.ca.gov, broken down by year, topic, publisher and use. The final section asks how consistently the two portals federate the same datasets.
By year first published
Year the dataset was first published on either portal (harvested copies keep the upstream issue date).
By topic
Keyword classification of title, tags and description. Method
By publisher
Page views are CKAN dataset-page visits, all time and last 14 days.
Most-viewed datasets
Federation: one dataset, two catalogs
AB 1755 relies on federation: the same dataset should be findable, and described the same way, on both portals. The portals harvest each other. DWR, CDFW and Ocean Protection Council records flow from CNRA to data.ca.gov, and Water Board records flow the other way through a harvest source CNRA names "AB1755 - CA Water Boards." These charts test how well the copies line up. They ignore the filters above, except Publisher.
Is the copy also filed under Water?
Federated pairs where at least one copy is in a Water group.
Where the views land
Share of page views on the origin portal vs. the harvested copy, federated pairs.
Harvesting resets the clock
data.ca.gov copies harvested from CNRA: year first published vs. year the copy was (re)created on data.ca.gov.
What feeds the CNRA Water group
Records by harvest source. Sources whose names start with "AB1755" are the statute's federal-data harvests.
Method and caveats
- Data: CKAN API on both portals (
package_search,package_show?include_tracking=true), fetched . Full table: water_datasets.csv. Rebuild withpython3 scripts/build_water_dashboard_data.py. - "Published water dataset": a record in the portal's Water group. In "Both" mode, a federated pair counts once, and its views are the sum across both copies.
- Federation match: exact, through harvest metadata (the harvested copy's
guidequals the upstream record's id). Unlinked records are also matched on title and publisher. - Views: CKAN's built-in tracking counts unique visitors per day on the dataset page. It does not count API calls, direct downloads from external URLs, or ArcGIS service hits, so heavily automated datasets are under-counted. The tracking start date differs by portal and is not published, so compare views within a portal more confidently than across portals.
- Topic: the first match in an ordered keyword list (see
TOPICSin the script). Each dataset gets exactly one topic, and edge cases will be imperfect. - Publisher: DWR, State Water Board, CDFW. "Federal" covers federal agency organizations and CNRA's Water Data Partners harvest. "Other" includes the Ocean Protection Council, BCDC, the Delta Stewardship Council and others.