The Aga Khan Visual Archive is a rich collection to target for this, but do not provide an easily accessible interface to grab the records. Need to contact folks at MIT to see if we can get a dump of these records. Otherwise, develop a crawler that starts at

https://dome.mit.edu/handle/1721.3/45936/browse?type=dateissued

which gets records like this:

https://dome.mit.edu/handle/1721.3/37656?show=full

This is a DC metadata with thumbs at

https://dome.mit.edu/bitstream/handle/1721.3/37656/126635_cp.jpg?sequence=1

Dome is written on DSpace, so hopefully there's at least OAI

Get the entire collection of movie posters from Princeton beyond the sample MODS records already in the repository.

There's an error in the expanded Penn Museum Egyptian CSV in processing for DLME:

CSV::MalformedCSVError: Unquoted fields do not allow \r or \n (line 2).

Spatial data was taken from a query of a bounding box of the MENA region and includes some items that are not particularly interesting to this endeavor. Go through and weed out vector datasets.

Follow up with QDL on AV material (http://av.qnl.qa/)

Harvard hosts their spatial data in GeoPortal (http://pelham.lib.harvard.edu:8080/opengeoportal/) for the FGDC. Unfortunately the metadata links they provide (e.g. http://hgl.harvard.edu:8080/HGL/jsp/HGL.jsp?action=VColl&VCollName=G7460_1853_H8) don't actually open to the layer (I suspect an OpenGeoPortal update broke the links). Need to investigate what's going on.

Obtain records from Aga Khan Documentation Center for indexing

The Department column in the downloaded file has the department field set as "Islamic". This needs to be updated to be "American Numismatic Society" as part of the download/massage process.

We need to get the endpoint for Princeton's IIIF collection for indexing.

http://vcoai.lib.harvard.edu/vcoai/vc?verb=ListRecords&set=ihp&metadataPrefix=mods

Contact Ethan for access to the metadata for the Islamic Department

Right now the Map functionality comes from a spatial query from Stanford's GeoBlacklight instance (https://earthworks.stanford.edu/?bbox=28.959961+20.715015+47.592773+38.891033&per_page=100) and manually paging through maps within a bounding box (28.959961 20.715015, 47.592773 38.891033). Specific maps are then "found" (with mechanize and added to an array of ids to look up in the OpenGeoData layers information.

This is an unscalable system that will break.

We need a system that will do the following:

• Work with a group to identify appropriate bounding box for maps
  • Include time-boxed expansions (e.g. Italy, Spain, Japan, etc.)
• Create a more efficient method of retrieving pointers to resources than looking up the JSON identifiers
• Method for synchronizing the metadata records between institutions