Can we infer an approximate leaf count from data we already have?

[regineheberlein]

User story

Following a conversation with Dan L.: We would like a better way to estimate the cost of digitization. The cost of digitization is by number of images, but archival description being by its nature aggregate-level, not item-level, we never have a page count available to create an estimate from, and counting pages for every digitization project is not practical because we frequently digitize many boxes at a time.

Dan is wondering whether there is a way to infer an approximate leaf count from the data we already have:

• can we get image counts for already-digitized collections from the manifests
• can we get folder counts for already-digitized collections from ASpace
• can we get box size from ASpace
• with those three data points, can we infer from the already-digitized collections an average image count per box size and/or folder count

Variables include:

• single- vs. double-sidedness
• carrier volume may fluctuate significantly (e.g. onion skin v. photo stock)
• more folders add bulk--can we benchmark folder volume

Questions that may or may not have a bearing on this include:

• would it help if we added box/folder/paper weight as a data points
• can we make predictions on the carrier based on the nature or period of the collection

Resource record URI's to pull this data from:

TBD

Fields to include:

sum of folder counts for all associated top_containers; possibly do id's?; cid's to get manifests with