This document discusses how to provide content to Europeana Fashion (EF), including:
- The EF metadata schema called EDM-fp based on Europeana Data Model
- Providing metadata using an Excel template, CSV export, or XML export and mapping it to EDM-fp using the Mint tool
- Mapping values to the EF Fashion Thesaurus for object types, techniques, materials, colors, and subjects
- Uploading media files either to the EF repository or linking to files on a content provider's own webserver
- Common issues encountered like unstructured data, invalid links, and missing required fields are also covered.
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Submitting content to Europeana Fashion
1. EUROPEANA FASHION
Providing Content to EF
Providing Content to Europeana Fashion
Using the Europeana Fashion platform:
Tools, best practices and use cases
Antwerp, February 25th, 2015
Presentation made by <name here>
Co-funded by the European Commission
within the ICT Policy Support Programme
Henk Vanstappen
IMMD
Connected to:
9. EUROPEANA FASHION
Providing Content to EF
Europeana Fashion Datamodel
• Set of fields for fashion content
• Notation guidelines
• Based on the Europeana Data Model
(EDM)
• Connected with Europeana Fashion
Thesaurus
10. EUROPEANA FASHION
Providing Content to EF
Fashion Thesaurus
• Fashion related concepts
• Object types, techniques, materials, colors
• Multilingual
• Optimizes search results
• Enhances filtering of content
16. EUROPEANA FASHION
Providing Content to EF
Mapping metadata to EDM-fp
• Create dataset from export or template
• Upload dataset to Mint tool
• Select corresponding elements
• Use custom functions
• Transform your data to EDM-fp
17. EUROPEANA FASHION
Providing Content to EF
Mapping metadata to EDM-fp
• Create dataset from export or template
• Upload dataset to Mint tool
• Select corresponding elements
• Use custom functions
18. EUROPEANA FASHION
Providing Content to EF
Mapping values to Fashion Thesaurus
• Select corresponding values for:
• object type
• technique
• material
• color
• subject
19. EUROPEANA FASHION
Providing Content to EF
Mapping metadata to Fashion Thesaurus
• Select corresponding values for:
• object type
• technique
• material
• color
• subject
23. EUROPEANA FASHION
Providing Content to EF
Top 6 issues, pitfalls, problems, …
• Unstructured data:
• multiple values in one field
• no separate fields (e.g. color)
• Invalid links to media files
• Invalid CSV files
• Typo’s, trailing spaces, capitalization
• dc:type missing
• Dates must be YYYY-MM-DD or YYYY
24. EUROPEANA FASHION
Providing Content to EF
Documentation and support
• EDM-fp documentation
• EDM-fp template
• YouTube instruction videos
• Tools: Filezilla, XNView, LibreOffice
• Basecamp Content Ingestion Group
• Helpdesk
25. EUROPEANA FASHION
Providing Content to EF
Thank you!
henk@immd.be
Presentation made by <name here>Presentation made by <name here>Co-funded by the European Commission
within the ICT Policy Support Programme Connected to:
Editor's Notes
Let’s first look at the portal, probably you have seen this: in EF you can search 1000s of fashion objects, e.g. looking for Shoes
.. and as you can see: this content comes from 20+ institutions all over Europe
Each object has metadata – some rich, some poor – I a structured way
so the question is: how do we aggregate all this content and make sure users can search all this – in different languages.
The complexity is not really in the images, but in aligning the metadata so it is uniform and searchable
To achieve this we developed a metadata schema. This is the same record as the previous slide, but presented in machine readable format. This is the information that is fed to the portal, and here you can see the ‘real’ names of all the fields that are shown on the portal.
The schema we developed is called EDMfp: a profile of EDM specifically for fashion content.
the thesaurus enhances multilingual searching and filtering of content, as you can see here for color (it’s not filtered yet)
So how is this done? What do you need to do if you provide content to EF and fit your data into this schema and thesaurus?
There are 2 possibilities: the first is when you don’t have a database with metadata (yet). In that case, you can use the schema to describe your collection. You can use an excel template we made and encode directly in EDM-fp. As you can see, the column headers have the same name as the EDM-fp elements.
But of course many institutions have metadata ready and can export data from their database, according their own schema. This may be a CSV export…
… or an XML (or JSON).
You can choose whatever suits you best: CSV is quite simple to create, but XML is better for more complex data.
As you can see, this export is according to the schema of the institution’s database.
Either way: the dataset has to be ingested then in the Europeana fashion database, in the following steps.
(the next presentation will go into this in more detail)
On the left you see the elements of your own database export. Here you can tell Mint that – for example – Tipologia’ corresponds to ‘dc:type’ element of EDM-fp.
If you’ve used the Excel template, it’s straightforward to do this.
When elements are mapped, values from the thesaurus need to be mapped too.
…while here you tell mint that – in a given element dc:type - the literal value ‘Sandalo’ corresponds to the ‘Sandals’ concept of the EF Thesaurus.
Or you might define that ‘bordeaux’ in the dc:color element corresponds to the concept Red, etc.
Next, you’ll need to provide the images (or videos, or PDF files)
Media files can be images, video or text documents. It’s also possible to submit multiple media files per record.
Again, there are several ways of doing this:
EF provides a ‘repository’: a server space where you can put your image files (not to big, preferably in jpg format).
You only need to make sure the file name corresponds to a metadata element in the dataset you provided.
If your media files are on line (on a webserver), you can alternatively provide a direct link. The portal will then fetch the media from there.
So both options are valid, but make sure the the link is persistent. If this is not possible, better upload a copy of your files to the repository.
This is the scary slide: things that will go wrong…