Date ranges based on non-collecting events

[ekrimmel]

At CHAS we have a significant portion of our malacology and entomology collections where the collecting date is unknown and was likely never recorded, but where we do know the accession date and/or collector. I am wondering how/if other collections have tried to incorporate non-collecting events, such as the receiving date or a collector's death, into their specimen records.

For specimens without a date my standard when migrating them into Arctos has been to assign a range with an unlikely BEGAN_DATE (1800) and an ENDED_DATE equal to whatever is today's date. In our other collections no date is the exception rather than the norm. However, no date CHAS entomology and malacology specimens are the norm rather than the exception, and we particularly care because the majority were collected pre-1920. Once you get a century back granularity to days or even months or years gets less important.

I could assign date ranges with an unlikely BEGAN_DATE (1800) and an ENDED_DATE equal to the date received in the accession, or if that is unknown then a BEGAN_DATE equal to the collector's birth and an ENDED_DATE equal to the collector's death. We have this level of data available for our prevalent collectors. My concern is three-fold: (1) how to document where these date ranges came from, especially considering how our data look once they get pushed to aggregators; (2) how to do this with the most consistency; and (3) how to do this with the most longevity, e.g. if an accession date was entered incorrectly and later we fix it, can that fix get carried over to a new specimen event date for specimens connected to that accession?

Thanks for any thoughts...!

[ebraker]

Hi Erica, Below is our policy. We also use 1800-01-01 as a default and then make a standard note in collecting event remarks. We don’t have a way to globally fix these should a date error/resolution be detected down the road other than sifting out records linked to a given accession or collector one by one. Best, Emily Unknown Dates Some specimens do not possess date information. However, an informed attempt at narrowing unknown dates to a more precise range greatly enhances a specimen’s research value. Clues such as collector life span, expedition logs, associated specimens, previous institution numbers/records, trade tags, and donation dates can help home in and bound an unknown date range. * If absolutely no additional clues can be inferred as to when a specimen was collected, the collecting date is entered as: Verbatim Date = [no verbatim date data] Begin = 1800-01-01 End = current date (as the specimen in hand could be collected no later than the present). In capturing historic specimen/lot data, any additional clues such as Received date for materials gifted to the museum or the timestamp for when the specimen was first entered into previous databases such as Filemaker can serve as the upper boundary. * If supplemental pieces of information aid in deriving a collecting date range, such sources must be noted in the Collecting Event remarks (i.e., "Date range is approximated. Date range boundaries were estimated using a default began date of 1800 and [Collector's DOB/DOD, specimen donation date, date cataloged, etc.]." From: Erica Krimmel [mailto:notifications@github.com] Sent: Thursday, November 30, 2017 3:36 PM To: ArctosDB/arctos <arctos@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: [ArctosDB/arctos] Date ranges based on non-collecting events (#1339) At CHAS we have a significant portion of our malacology and entomology collections where the collecting date is unknown and was likely never recorded, but where we do know the accession date and/or collector. I am wondering how/if other collections have tried to incorporate non-collecting events, such as the receiving date or a collector's death, into their specimen records. For specimens without a date my standard when migrating them into Arctos has been to assign a range with an unlikely BEGAN_DATE (1800) and an ENDED_DATE equal to whatever is today's date. In our other collections no date is the exception rather than the norm. However, no date CHAS entomology and malacology specimens are the norm rather than the exception, and we particularly care because the majority were collected pre-1920. Once you get a century back granularity to days or even months or years gets less important. I could assign date ranges with an unlikely BEGAN_DATE (1800) and an ENDED_DATE equal to the date received in the accession, or if that is unknown then a BEGAN_DATE equal to the collector's birth and an ENDED_DATE equal to the collector's death. We have this level of data available for our prevalent collectors. My concern is three-fold: (1) how to document where these date ranges came from, especially considering how our data look once they get pushed to aggregators; (2) how to do this with the most consistency; and (3) how to do this with the most longevity, e.g. if an accession date was entered incorrectly and later we fix it, can that fix get carried over to a new specimen event date for specimens connected to that accession? Thanks for any thoughts...! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#1339>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AKz7JTnHXh-bKTqYwCH7caznoMYZ1MHNks5s7y26gaJpZM4QxTt1>.

[dustymc]

how to document where these date ranges came from

Collecting event remarks seems correct.

how our data look once they get pushed to aggregators

#1291 (comment) - GBIF is adding crazy precision to some dates, and I generally trend towards "1800" rather than "1800-01-01" - the range/end result is exactly the same, but I think the year-precision start/stop dates do a better job of alerting users to the situation.

I also usually put something like "before {today}" in verbatim_date for the same reason - it's not really verbatim, but I think it does help get the idea across to users.

Someone should add whatever comes of this to http://handbook.arctosdb.org/documentation/collecting-event.html#verbatim-date

can that fix get carried over to a new specimen event date

I don't see a mechanism for automating that. Perhaps you could leave remarks in the "source" data, but that's obviously far from perfect.

[dustymc]

Happy?