...
- Implemented "QAQC" in scalar data products. Data will start to be analyzed for invalid values and flagged. It will take some time before all instrument data is being analyzed, but the infrastructure is now in place to support it.** Data can be analyzed "manually" (by hand) or automatically by software.
- Automatic tests include (not an exhaustive list) out of range tests, spike tests, high gradient tests, and stuck value tests.
- Scalar data products (CSV, MAT, PNG, PDF) now show a QAQC flag for each record. CSV and MAT list the flag value, ranging from 0 to 9.PNG and PDF plots highlight potentially invalid data with yellow or red colours, and with a yellow or red flag in the center of each group of flagged data.
- In a later release the user will have the ability to select "clean" or "raw" data in their data products. For now all non-resampled data is "raw", meaning that invalid data is still present but is flagged. All resampled data is "clean", meaning that invalid data has been removed; for the CSV NaN option it will be replaced with "NaN" (not a number); for the other options it will be removed entirely.
- In the case of averaged data, if less than 80% of the expected data values for a given averaging point are not present (either due to data gaps or failing a QAQC test), that averaged point is treated as a failed QAQC data point.
Anchor 20130321 20130321
...