Linked Data at the BBC - Speaker Deck

Slide 1

Slide 1 text

Linked Data at the BBC Paul Rissen, Senior Information Architect, BBC Future Media Thursday, 16 August 2012 Hello.

Slide 2

Slide 2 text

A little Web Theory... Thursday, 16 August 2012 First, a little bit of Web theory and history.

Slide 3

Slide 3 text

THE WEB DONE RIGHT THE DREAM OF SIR TIM BERNERS-LEE http://www.ﬂickr.com/photos/agirregabiria/5731646386/ Thursday, 16 August 2012 What is linked data? what is the semantic web? it’s ‘the web done right’ - it’s what TimBl ﬁrst proposed.

Slide 4

Slide 4 text

IN THE BEGINNING... RISE OF THE MACHINES http://www.ﬂickr.com/photos/marc_smith/6246433245 Thursday, 16 August 2012 At ﬁrst, it was all about the machines - you had to know the exact address of a machine.

Slide 5

Slide 5 text

WHERE WE ARE NOW... THE WEB OF DOCUMENTS http://www.ﬂickr.com/photos/seanvenn/3687559590 Thursday, 16 August 2012 Now, it’s about the documents - the addresses we type, mean we retrieve documents - no matter what computer they’re on.

Slide 6

Slide 6 text

IT’S NOT THE DOCUMENTS, IT’S THE THINGS... THE GIANT GLOBAL GRAPH http://www.ﬂickr.com/photos/blmurch/522427911/ Thursday, 16 August 2012 But really, people (mostly) don’t care about the documents. They care about the concepts discussed. And that’s what linked data is. The data part is just how we represent it - what matters is the concepts.

Slide 7

Slide 7 text

WHAT’S THE MEANING OF THIS? PEOPLE CAN EXTRACT IT, MACHINES, ON THEIR OWN, CAN’T. http://www.ﬂickr.com/photos/46871088@N00/3973599885 Thursday, 16 August 2012 Why? Because as humans, we can read a document and construct a web of meaning in our heads. Computers can’t. They’re reduced to text string matching. But we can help them get smarter. And by doing so, we can get better results.

Slide 8

Slide 8 text

THE URI IS THE THING THREE BASIC RULES Thursday, 16 August 2012 Use URIs to identify things, not just pages. Use HTTP URIs so people can ﬁnd those things using web browsers. Link to other relevant URIs so people can discover more things.

Slide 9

Slide 9 text

Putting this into practice Thursday, 16 August 2012

Slide 10

Slide 10 text

BBC PROGRAMMES A PAGE FOR EVERY PROGRAMME Thursday, 16 August 2012 page for every programme, ﬁndable in google, we can aggregate blogs/tweets about a programme.

Slide 11

Slide 11 text

BBC MUSIC A PAGE FOR EVERY ARTIST Thursday, 16 August 2012 page for every artist played across BBC radio - we can pull people through from shows > artists and back.

Slide 12

Slide 12 text

BBC NATURE A PAGE FOR EVERY SPECIES Thursday, 16 August 2012 Vast archive of David Attenborough etc - page for every species - search for ‘Lion’ or ‘Polar Bear’, and we’ll be near the top.

Slide 13

Slide 13 text

User Beneﬁts Thursday, 16 August 2012

Slide 14

Slide 14 text

PERSISTENCE WEB DESIGN FOR THE FUTURE Thursday, 16 August 2012 Persistence is important, because it’s terrible user experience if you lose pages. It’s the main reason you’ll be hated.

Slide 15

Slide 15 text

MULTIPLATFORM DESKTOP, MOBILE, TABLET, IPTV - ONE WEB Thursday, 16 August 2012 Doing things this way makes multiplatform easy. Rather than having to invest loads of money on each new platform that comes along, get the data, the information about the things right ﬁrst, then just build presentation layers.

Slide 16

Slide 16 text

OUR WEBSITE IS OUR API Thursday, 16 August 2012 We want to encourage others to build stuff with our stuff.

Slide 17

Slide 17 text

Thursday, 16 August 2012 ...stitching everything across the BBC into one single thing. Because that’s how our minds work. It’s a Web, not just a delivery platform.

Slide 18

Slide 18 text

Thursday, 16 August 2012

Slide 19

Slide 19 text

Thursday, 16 August 2012

Slide 20

Slide 20 text

Business Beneﬁts Case Study: Sport Thursday, 16 August 2012

Slide 21

Slide 21 text

TOO MANY MANUALLY MANAGED INDEXES Thursday, 16 August 2012 Currently there are too many pages to manage manually for a small editorial team 7 days a week We need to drastically reduce the number of pages that are manually managed - by approximately 300 There are approx 320 manually managed ‘indexes’ - 150 or so in football, rugby teams, cricket teams and further ambitions around F1 Difﬁcult to integrate statistical information with editorial which is why the metadata and tagging projects are an integral part of long term strategy

Slide 22

Slide 22 text

WORLD CUP 2010 A PAGE FOR EVERY TEAM, GROUP AND PLAYER Thursday, 16 August 2012

Slide 23

Slide 23 text

THE SPORT ONTOLOGY Thursday, 16 August 2012 We’re seperating concepts from content. Using the intelligence we have about the concepts to make the UX better - and lighten the workload for our staff. Does need good management of the data, but the beneﬁts are worth it.

Slide 24

Slide 24 text

DYNAMIC SEMANTIC PUBLISHING Thursday, 16 August 2012

Slide 25

Slide 25 text

DYNAMIC SEMANTIC PUBLISHING Thursday, 16 August 2012

Slide 26

Slide 26 text

DYNAMIC SEMANTIC PUBLISHING Thursday, 16 August 2012

Slide 27

Slide 27 text

LONDON OLYMPICS 2012 Thursday, 16 August 2012 Working with external sources & partners too - geonames, LOCOG data etc. Stitching BBC into the wider Web.

Slide 28

Slide 28 text

IN SUMMARY, THEN.. - THE WEB CAN BE ABOUT MORE THAN JUST DOCUMENTS - DESIGN AROUND THINGS YOUR USERS CARE ABOUT - ONE URI PER THING - PERMANENCE IS KEY - DESIGN FOR THE WEB FIRST, THEN APPROPRIATE PLATFORMS - LINKING CONCEPTS PROVIDES GENERATIVE JOURNEYS - SAVING LABORIOUS, HAND-CRAFTED LINKING WORK - STITCH INTO THE WIDER WEB - BENEFITS FOR YOU, BENEFITS FOR ALL Thursday, 16 August 2012

Slide 29

Slide 29 text

Thanks! Any Questions? Paul Rissen, Senior Information Architect, BBC Future Media @r4isstatic http://www.r4isstatic.com Thursday, 16 August 2012 Hello.