The document describes the birth of the Newsroom of Things project at the BBC. It details how the Juicer tool was created to automatically tag BBC news content with linked data tags. This allowed the BBC to test using real news data without extensive manual tagging. Over time, the Juicer was improved and scaled up to include more data sources. It is now used to explore opportunities at the intersection of technology, journalism, and data through projects like analyzing media coverage of events. The document concludes by outlining some future directions for the Newsroom of Things, including expanding linked data workflows and exploring partnerships.
8. So, my focus
ARRGHHHH VALUE
Near Term Long Term
ONLY
LOOK
HERE !
It drives
best
ROI
It opens
future
opportunities
It helps create
productive
teams
15. “How might BBC News
use Linked Data
to connect and
contextualise
our News Stories?”
16. Answer: The Juicer
• before we change the Newsroom?
• and without the need for MAGIC DUST?
So – how do we explore this
1
Get
Content
2
Extract
Concepts
3
Match to
DBpedia
4
Annotate
Content
5
Push to
Triplestore
6
Expose
via
API
18. OK. In other words:
The Juicer automatically
tags News content with
Linked Data Tags
19. ...& it allows us to test ideas,
using real data,
without an army
of tagging experts.
(NB it is not as good as real humans, yet)
20. The Timeline of
The Birth of THINGS*
in BBC News
* “Things” being: Linked Data, real world concepts (or “tags”)
that are understandable by machines.
21. Window on the NEWSROOM
Production App
Journalists
doing Manual Tagging
Storyline
2013 2014 2015
Summer
Labs Added “Non BBC” Sources
Start
App prototype PAIN
Sources ++
YES!
R&D
DISCUSS
A LOT
FUNDING
CRISIS
Joined
Connected
Studio
What’s the
editorial
story L.D.
concept?
v1
Conceptual
Struggle
Sceptical
Of Linked
Data
explore
APP
“Yes – it
will work!”
Innovation
Support
Production
Scaling
Vote
2014
Used
Linked
Data
This saved
us!
Thanks
to Robin
JT
Joins
JUICER POC in
the Newsroom
22. And so “The Things”
were born in BBC News.
And launched to our users
via BBC mobile Apps
24. What did we find?
1. A Prototype is worth 1000 meetings
2. There will be pain
3. Loose coupling of teams/services is key
4. Keep referring to point 1
35. NB – the tag associations here
are “co-occurrences”, and so
there is no indication of what the
co-occurrence means.
i.e. It could be negative, positive,
or non associative.
All we know is that the terms
have co-occurred in News
Articles in our source set.