Select a published dataset, narrow the columns and rows, and download the result as a CSV or Parquet file.