The data should be closely inspected to check that all data that should be there are there and that the dataset looks healthy. This is best done using the (potentially many hundreds) of plots that will have been created during the user pipeline, but there will also be useful information in the letter sent by the JIVE correlator support scientist when the data were released. This is also a good point at which to start flagging the data. Copy the FG table assembled during the pipeline. Additional flagging will almost certainly be required.