Tracking Data

Latest revision as of 11:04, 6 January 2014

This page refers to the current procedure for recording administrative metadata during the capture process. For older processes of recording tracking data, see TrackingFiles.

[edit] Procedure

[edit] Integrate the tracking data columns before capture begins

1. Open the collection's metadata spreadsheet
2. Open trackingcolumns_template.xlsx (found in S:\Digital Projects\Organization\Digital_Program_Logs\TrackingFiles\TrackingFiles_database_files)
3. COPY tracking columns to the end of the metadata spreadsheet or enter these column headers manually, just as they are given here:

    Number of Captures
    Captured with
    Captured by
    Date
    OCR? (1=yes or 0=no)
    DS Notes
    Metadata changed

4. Save the metadata spreadsheet (and close trackingcolumns_template.xlsx)

[edit] Use the tracking data columns during capture

• Include data for items that you actually capture, not things you skip (see Skipped Items for how to track those)
• If you begin but don't finish an item (even if you're planning to finish it soon), note that in the DS Notes column; this especially important for collections that multiple people are working on, or that you're going to put on the backburner for some reason -- basically, make sure someone else picking up the collection will know exactly where you are, even if you're in the middle of an item
• Check your Number of Captures against the Format column in the metadata part of the spreadsheet -- the metadata should be made to match the number of captures you make (if there's a big discrepancy, see Procedural Anomalies)
• Be sure to fill in the OCR? column so that Tesseract can do its job

[edit] Remove and save the tracking data columns

• This happens after the spreadsheet has been batched, if necessary
• COPY Filename column from the metadata spreadsheet to a new sheet
• MOVE Tracking Data columns to new sheet
• Make sure you also copy the column headers
  • If you don't, you'll get an error on upload -- because the script automatically assumes the first row is the header, it will interpret that your first item in Scans hasn't been entered into the log
• Save new sheet as tab delimited .txt file [collNum.batchNum.log.txt] in the collection's Admin folder
  • Example: u0001_2007001.25.log.txt

[edit] Rationale for Change

[edit] Problem

In December of 2012, it was proposed that we rethink the current tracking model, for two reasons:

1. Numerous fields in the TrackingFiles document were not being used, and they were overburdened with validation rules
2. TrackingFiles documents were separate files, housed in a totally separate location from the rest of the collection's documentation

[edit] Advantages of New System

We then proposed the simplified and integrated tracking data procedure outlined above. This process has many advantages for our workflow:

• There is no longer a need to keep two spreadsheets open during capture, which improves our data collection
• With everything in one spreadsheet, it is easier to
  • check the metadata against the tracking data, which allows for more accurate recordkeeping
  • isolate batch tracking data (as the entire working metadata/tracking spreadsheet is being "batched")
• The tracking data has been pared down to essential fields but is open to future reconfiguration if necessary

[edit] Effects and Potential Effects

The process has the following impacts on workflow:

• No added step at preparation for capture, just a rewriting of the old one (adding template columns to existing sheet rather than saving template sheet as new document)
• No added step at preparation for upload, just a reworking of the old (transferring and saving the data rather than simply saving a spreadsheet)
  • admittedly, the copying/removing procedure introduces risk, but it is a risk we are accustomed to taking: we have always done something similar to extract the OCR list from TrackingFiles
  • if the columns are not removed, Archivist Utility will point them out during the MODS-making process, and the problem can be corrected
• The new model contains most of the same data that was effectively used in the old TrackingFiles documents, and can be named and archived in the same way
• Done properly, the new model should not impact the Metadata Unit at all; done poorly, it will not break our current system (columns will be mapped to MODS in a way that doesn't trigger a fatal error)
• Legacy collections will not require manual migration to the new system
  • There was nothing inherently broken about the old system, so it can continue where it needs to
  • New form tracking can be instituted as outlined above at any point in ongoing collections because we document them in batches