It's all about data! Media Management & Information Management at NISV (MANDERS and VAN ARKEL)
1. It’s all about data!
Media Management
&
Information Management @
The Netherlands Institute for Sound and Vision
Tim Manders tmanders@beeldengeluid.nl
14. Information Management House blueprint I
Information
Governance
Information Architecture
Data Quality Reference &
Master Data
Data
Warehouse
Metadata Assets
Data about Data
Information Security Management
People
Technology
16. How do we stay compliant & relevant?
Who does what, when, how & why?
How to get good
(enough) data?
What are the definitions that make us understand our data?
How do we secure (access to) information?
People
Technology
How to achieve
(open) access &
interoperability?
How to make
informed decisions
based on analyses?
How to maintain
& support databases
& records?
How are the files
managed?
Information Management House blueprint III
17. How do we stay compliant & relevant?
Who does what, when, how & why?
How to get good
(enough) data?
What are the definitions that make us understand our data?
How do we secure (access to) information?
People
Technology
How to achieve
(open) access &
interoperability?
How to make
informed decisions
based on analyses?
How to maintain
& support databases
& records?
How are the files
managed?
Media Management rooms in house
Tim / Since a few years I’m part of Optimisation team within the Archive
Optimisation team since 2017 is part of a new information management department that has to deal with all information, not just program descriptions
We are moving into a NEW house with more furniture than only descriptive metadata.
Second part of presentation = tour through new Information Management mansion (under construction)
Look back at our last 3 FIATs world conference presentations!
Instead of structuring metadata manually, post-ingest, we embedding our thesaurus in broadcasters source system for persons
Result: uniform thesaurus labels that pinpoint relevant access point into collection items and also make these items linkable to other items, collections withing or outside our institute.
introduced the work of the new department of Optimisation Media Managers, bridging two gaps.
First the gap between the production environment and the archive.
We agreed with the broadcasters which information they have to supply at ingest to make assets findable. Title, summary, cast/crew
2. Second the internal gap between the prototype world of R&D and operational heart of the archive: Together we hooked up speaker labeling and term extraction to our MAM and apply it on our daily ingest. Now we are working on implementing facial recognition.
In 2016 then, we presented dreams about production metadata and automatically generated metadata validating each other.
Media Management in a nutshell has 4 ingredients1: If you cannot automate it, don’t do it yourself if you don’t have to (and metadatais created earlier on in production-archiving chain)
2: Use a thesaurus to structure metadata (makes the metadata uniform + linkable)
3. Automatic annotation, for two reasons.
1. To fill metadatagap
2. To achieve more finegrained access on a segment level
4. Combine efforts: concretely use production metadata to validate automatic metadata. Or combine different techniques.
Metadata is complete but we should check assumptions on needed fields on analytics on actual search behaviour
House ]currently is just a blueprint
We use it to check what aspects according to us define information management
What areas we can tackle as a department and what areas we have to share with others
or totally have to rent out rooms in the house because there is more expertise elsewhere
what are new areas of knowledge we have to explore, grow and professionalise
House is walled in by people and systems
It’s flooring plan is divided into a few layers:
On top the overarching roof of information governance. This is where is formalised how policies and strategies are being created or updated.
The information architecture layer is the level where processes & models are designed according to the policies & chosen standards that trickle down from the roof
Then we see a layer with multiple rooms: core activities / The heart of operation
The layer ‘data about data’: there where meta-definitions (glossaries or dictionaries) are being safeguarded. We need them to understand our data and it’s context better
The information security management layer in the end: that’s where it is arranged if, where and when information can be accessed and by whom, and ounder what conditions.
MM involves itself mostly with two blocks in the operational heart of the house:
quality of metadata in terms of accuracy and completeness
and strucuturing metadata via the thesaurus to make it uniform and linkable
New for our department is mostly the data warehouse block. We are exploring what we can do with business & customer intelligence tools and analyses.
An advantage of this framework also lies in the visualisations of the vertical connections: need to constantly check impact of new developments on all level