PrevIQ Docs
Open app

Data & freshness

Where PrevIQ's data comes from, how current it is, and how new exports are added.

The Pipeline area shows the operational state of your data: how much is loaded, when it last updated, and how new exports get in.

Database status

A set of indicators reports the live state of the data service:

Indicator Meaning
Normalized records Total organism-result rows currently loaded.
Prevalence rollups Number of pre-computed prevalence rows.
Database size On-disk size of the dataset.
Vendors Laboratories currently represented.
Last ingestion When data was most recently added (shown in your local time).

How data gets in

New laboratory exports are added by ingestion. PrevIQ accepts:

  • Ignite-style CSV exports, and
  • Excel templates/workbooks (.xlsx, .xlsm) that contain a Raw Data sheet (or whose first row begins Date, Accession, Panel, Positive).

  1. Provide the file

    Upload the raw export (or send it to your OSPRI contact for bulk loads).

  2. Identify the lab

    A vendor ID is required. PrevIQ reads it automatically from Ignite-style filenames (e.g. AFC-RPT-...); for other files you supply it.

  3. Parse → normalize → compute

    PrevIQ parses the rows, normalizes organisms and panels, stores the records, and rebuilds the prevalence rollups.

Note

Ingestion is idempotent. If you submit the same file twice, PrevIQ recognizes it by content hash and skips it rather than double-counting — so re-sending a file is always safe.

Data quality flags

Each ingestion run records QA flags so you know if anything unusual was seen, for example:

  • UNKNOWN_RESULT — a result value PrevIQ didn't recognize was encountered.
  • ALL_NA_ROW — a specimen whose every result was "not tested" (N/A).

Ingestion history

A history list shows past runs — vendor, source file, rows ingested, rows skipped, rows de-duplicated, and any QA flags — so you can confirm a load landed as expected.

Tip

If a number looks off after a fresh load, check the most recent ingestion run's row counts and QA flags first. See Troubleshooting.