airmosaic-ai

PM2.5 station acquisition and preparation

English

Select original hourly publication records or a processed daily product before acquisition. This reference has no tool ID or downloader. Use the station method for provenance, concentration state and completeness, the source table for coverage, and the homogenized method for adjusted daily values.

Acquire

Define network, dates, pollutants, station scope, product type and destination. Public routes with historical checks on 2026-10-26:

Quotsoft single-day CSV: 2024-01-01 station archive, hourly PM2.5 reshaped to long form; time-zone/hour-interval semantics and historical station coordinates remain unresolved.
PANGAEA homogenized release: 2015–2019 release and one seven-day series sample; series IDs are not confirmed national station codes.

Both require no account/API key. Follow the selected example for actual HTTPS/file handling and bounded requests; neither is a run-tool entry. Preserve original files, source/version/license, acquisition time and checksums. Review the original request state after failure rather than silently switching providers.

Prepare and check

Verify the actual schema and missing symbols. Select hourly PM2.5 explicitly, excluding PM2.5_24h and AQI. Reshape with string station IDs and retain original date/hour labels.
Define unique source/network/station/time/observation-type keys; report conflicting duplicates. Count finite, valid, unique observed intervals rather than rows.
Join station history by explicit ID and validity period; check uniqueness and unmatched rows. Unknown historical coordinates remain unknown.
Resolve time zone, hour start/end and day assignment before aggregation. Use the selected dated completeness policy; under the GB 3095 20-hour daily rule, keep insufficient days missing with counts and reasons. Do not count imputation toward observed hours.
Deliver long tables, mappings, station-history version and duplicate/missingness/coverage summaries using the preparation method. Existing daily homogenized products do not pass through hourly aggregation.

Store plans in AIRMOSAIC_CONFIG_DIR/pm25-stations/ and task artifacts under AIRMOSAIC_ASSETS_DIR/results/<project_id>/<run_id>/03_atmospheric_transport/pm25-stations/, separating raw, processed, metadata and quality. Reviewed reusable data use data/03_atmospheric_transport/PM25_Stations/<source>/<version>/. Keep authentication separate in AUTH_DIR when a selected provider requires it. Historical scripts require manual adaptation; see legacy interface notes.

This site is open source. Improve this page.