From Search to Discovery in our Future Library From Search to - - PDF document

From Search to Discovery in our Future Library

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search

Known Specific Focused

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search

Known Specific Focused

Discovery

Unknown Vague Open

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search Discovery

Locating a known item

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search Discovery

Locating a known item Browsing items by a known author

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search Discovery

Locating a known item Browsing items by a known author Browsing items similar to a known item

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search Discovery

Locating a known item Browsing items by a known author Browsing items similar to a known item Browsing items curated by a person you’re interested in

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search Discovery

Locating a known item Browsing items by a known author Browsing items similar to a known item Browsing items curated by a person you’re interested in Browsing items by physical appearance

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

The Continuum

Search Discovery

Locating a known item Browsing items by a known author Browsing items similar to a known item Browsing items curated by a person you’re interested in Browsing items by physical appearance Browsing random items

W e see a spectrum of difference between Search & Discovery, Search being more targeted and specific and Discovery being more open and vague. This slide also illustrates some examples of the different types of needs of users and how they might vary along that spectrum.

Why Discovery Is Important

At UTS Library we are implementing an Automated Storage & Retrieval System that will be located underground and adjacent to the future Library

• n Broadway. Currently the excavation is completed and the builder has

started on the underground vault which is to be completed by around September 2013.

Six aisles, 12,000+ bins: ~

• ne

million print items

After that, a Dematic ASRS will be installed over a number of months with six 15m cranes services the aisles and over 12,000 bins. In mid-2014 we will start loading this system with the first batch of our less well used print items from the Blake (City Campus) Library. The next load will come from the Kuring-Gai Campus Library at the end of 2015 when we merge both libraries in the City. W e estimate that the system will cope with annual relegations from the open shelves in the Library until well beyond 2040, thus keeping research collections on-site and within easy reach of Library patrons. With up to 80% of our print collection in this facility, we believe the investment is wasted if it isn’t used and to use it to its fullest extent, the items in it need to be discovered online. The serendipity of physical browsing has to be replaced in imaginative ways, online, by new features that excite curiosity and encourage requests from the facility.

How Discovery Is More Playful

W e think that successful Discovery is going to be more playful than our Search interfaces. So we are looking for inspiration beyond library applications and programs. And it isn’t a challenge to be solved by more benchmarking and best practice measurement. So, we’ve looked the the ways people are using mobile devices to discover new content when convenient to them. W e think there is a lot to learn from new content aggregators and discovery applications like Zite http://www.zite.com/ The UI is playful, intuitive and assists the application to learn more about your interests and needs. Other examples include StumbleUpon http:// www.stumbleupon.com/ and Artsy http://artsy.net/

How Discovery Is More Playful

Another example is how we are learning to discover new music by online platforms like iTunes, Pandora http://www.pandora.com/, rdio http:// www.rdio.com/ and Spotify https://www.spotify.com/au/ These online services assist users in exploring musical content well beyond the playlists

• f most radio stations or your own music library.

How Discovery Is More Playful

Other helpful discovery platforms like Pinterest http://pinterest.com/ and Urbanspoon http://www.urbanspoon.com/ or Y elp http://www.yelp.com.au/ have much to teach us about discovery too.

W e have even more help! In 2012 we started an Artist-in-Residence program and it continues in 2013. Our first Artist Chris Gaul focussed on looking at discovery in libraries in very different ways. As an artist he had a very different approach to the same challenge that we did as librarians and we found this to be extremely helpful in improving our understanding, providing a new perspective and many original ideas, like the discovery of text based content through sound.

How Our Artist In Residence Has Helped

W e have even more help! In 2012 we started an Artist-in-Residence program and it continues in 2013. Our first Artist Chris Gaul focussed on looking at discovery in libraries in very different ways. As an artist he had a very different approach to the same challenge that we did as librarians and we found this to be extremely helpful in improving our understanding, providing a new perspective and many original ideas, like the discovery of text based content through sound.

How Our Artist In Residence Has Helped

Libraries Rational Efficient Logical

W e have even more help! In 2012 we started an Artist-in-Residence program and it continues in 2013. Our first Artist Chris Gaul focussed on looking at discovery in libraries in very different ways. As an artist he had a very different approach to the same challenge that we did as librarians and we found this to be extremely helpful in improving our understanding, providing a new perspective and many original ideas, like the discovery of text based content through sound.

How Our Artist In Residence Has Helped

Libraries Rational Efficient Logical Artists Emotive Intuitive Questioning

W e have even more help! In 2012 we started an Artist-in-Residence program and it continues in 2013. Our first Artist Chris Gaul focussed on looking at discovery in libraries in very different ways. As an artist he had a very different approach to the same challenge that we did as librarians and we found this to be extremely helpful in improving our understanding, providing a new perspective and many original ideas, like the discovery of text based content through sound.

How Our Artist In Residence Has Helped

> Understanding beyond what we know > Fresh Perspective > New & Original Ideas

W e have even more help! In 2012 we started an Artist-in-Residence program and it continues in 2013. Our first Artist Chris Gaul focussed on looking at discovery in libraries in very different ways. As an artist he had a very different approach to the same challenge that we did as librarians and we found this to be extremely helpful in improving our understanding, providing a new perspective and many original ideas, like the discovery of text based content through sound.

From Search to Discovery

Discovery is not a system problem.

But Discovery is not just something that needs to be provided in an online

• environment. The people in libraries need to be involved and this means

both staff and users.

From Search to Discovery

Rediscovering the curatorial role.

research > acquire > arrange > describe > provide

W e think there is a need to encourage more understanding and practice of the entire curatorial role in libraries. Sometimes a focus on efficiencies and systems makes us lose sight of this role.

From Search to Discovery

CURIOSITY ENGAGEMENT DELIGHT SERENDIPITY SURPRISE

These are some of the things we need to aim at in being more active as curators of our collections. These things should be the objectives of our events and programs that are delivered within our libraries to assist the process of discovery, as well as our online programs and services.

... [technological] solutionism presumes rather than investigates the problems that it is trying to solve, reaching "for the answer before the questions have been fully asked."

Evgeny Morozov, To Save Everything, Click Here: The Folly of Technological Solutionism

From Search to Discovery

And whatever we do needs to start with a better understanding of our users and their needs and behaviours. W e’ve recently done a lot of user experience testing and research at the start of our projects so that the results deliver services users need and will use. W e are also looking at hiring and anthropologist or ethnographer to embed that research as an ongoing program within our Library. I used this quote deliberately to be provocative at THETA. I grabbed it from a long review

• f Evgeneny Morozov’s book in The Atlantic http://www.theatlantic.com/technology/archive/

13/03/toward-a-complex-realistic-and-moral-tech-criticism/273996/ W e have the book on

• rder, but I’ve not yet seen or read it.

I have since read this criticism of Morozov’s book by Cory Doctorow which is a bit of a worry: http://boingboing.net/2013/04/14/blowing-up-morozovs-to-sav.html Doctorow refers to Tim Wu’s review but offers further criticism and analysis of his own. I have heard Doctorow speak and read his work online regularly and respect his opinion on matters including this subject. This indicates that I may have been too eager to support Morozov’s analysis without fully digesting his argument. Regardless, the point of this slide it to emphasise that we start with understanding our users and not leap into following the latest fad or fashion and then delivering them something they will not use.

Discovery & Open Access Content

From Search to Discovery

Finally for my part, I’d like to briefly mention some of our recent efforts to assist others to discover the content created in our institutions, such as our Open Access journals. A UTSePress we publish 16 of these OA journals, but they could be more widely read, so we’ve contacted both The Conversation and BrowZine in an effort to provide them with an easy way to include OA content in their own services. Josh will cover more of the technical detail behind that initiative.

Josh Vawdrey

IT Services Manager

Speaker 02

Aggregation Recommendation

Where We Are Heading

From a technology perspective these are our two main objectives at present regarding Discovery.

Where We Are Heading

Simplified Discovery: Aggregation

Recently we have provided aggregated search at article level for our journal subscriptions via a Primo Central Index. In looking at Simplified Discovery we have taken some inspiration from relatively new and popular aggregators like the Huffington Post and Zite.

Aggregation: Article-level search

Primo Central Index Faceted browsing Custom, HTML5 frontend Article-level search http://articles.lib.uts.edu.au/search: Primo Central Index

• custom front-end with a focus on extensibility and integration
• RESTful web service
• HTML5 interface: jQuery, Backbone.js, Twitter Bootstrap UI

Non-traditional sources

• "Huffington Post model"
• enhanced discovery
• increased exposure

Aggregation: Non-traditional sources

Enhance discovery Increased exposure Huffington Post model Non-traditional sources

• "Huffington Post model"
• enhanced discovery
• increased exposure

Aggregation: Non-traditional sources

Non-traditional sources

• "Huffington Post model"
• enhanced discovery
• increased exposure

Aggregation: Non-traditional sources

Here are a series of animated slides illustrating how we can incorporate non-traditional sources into our catalogue search. It is based on a Huffington Post aggregation model and enhances the discovery experience via the Library website. It is all done through self-submission which is demonstrated as an example using something from Chris Gaul being added. To set this up we needed to go out and identify the types of works that were being generated but not aggregated in our community. Through the Libarry website we can offer them increased exposure of their work.

Aggregation: Non-traditional sources

Here are a series of animated slides illustrating how we can incorporate non-traditional sources into our catalogue search. It is based on a Huffington Post aggregation model and enhances the discovery experience via the Library website. It is all done through self-submission which is demonstrated as an example using something from Chris Gaul being added. To set this up we needed to go out and identify the types of works that were being generated but not aggregated in our community. Through the Libarry website we can offer them increased exposure of their work.

Aggregation: Non-traditional sources

Here are a series of animated slides illustrating how we can incorporate non-traditional sources into our catalogue search. It is based on a Huffington Post aggregation model and enhances the discovery experience via the Library website. It is all done through self-submission which is demonstrated as an example using something from Chris Gaul being added. To set this up we needed to go out and identify the types of works that were being generated but not aggregated in our community. Through the Libarry website we can offer them increased exposure of their work.

Where We Are Heading

Simplified Discovery:

This next series of animations displays (just via wire frame outline) where we are headed in terms of Recommendation. Users initially see a large single search box (like Google) and below it can see a reference to recommendations. After that we see a personalised home page with a basic dashboard offering recommendations much like Amazon. These include personal, curated and more general recommendations and then lead into social recommendations that may be curated with the context of the current record.

Where We Are Heading

Simplified Discovery: Recommendation

• n the library website.

• Student/Staff Number

This next series of animations displays (just via wire frame outline) where we are headed in terms of Recommendation. Users initially see a large single search box (like Google) and below it can see a reference to recommendations. After that we see a personalised home page with a basic dashboard offering recommendations much like Amazon. These include personal, curated and more general recommendations and then lead into social recommendations that may be curated with the context of the current record.

Where We Are Heading

Simplified Discovery: Recommendation

• n the library website.

• Student/Staff Number

This next series of animations displays (just via wire frame outline) where we are headed in terms of Recommendation. Users initially see a large single search box (like Google) and below it can see a reference to recommendations. After that we see a personalised home page with a basic dashboard offering recommendations much like Amazon. These include personal, curated and more general recommendations and then lead into social recommendations that may be curated with the context of the current record.

Where We Are Heading

Simplified Discovery: Recommendation

• n the library website.

• Student/Staff Number

This next series of animations displays (just via wire frame outline) where we are headed in terms of Recommendation. Users initially see a large single search box (like Google) and below it can see a reference to recommendations. After that we see a personalised home page with a basic dashboard offering recommendations much like Amazon. These include personal, curated and more general recommendations and then lead into social recommendations that may be curated with the context of the current record.

Where We Are Heading

Simplified Discovery: Recommendation

• n the library website.

• Student/Staff Number

This next series of animations displays (just via wire frame outline) where we are headed in terms of Recommendation. Users initially see a large single search box (like Google) and below it can see a reference to recommendations. After that we see a personalised home page with a basic dashboard offering recommendations much like Amazon. These include personal, curated and more general recommendations and then lead into social recommendations that may be curated with the context of the current record.

Where We Are Heading

• n the library website.

• Student/Staff Number

This next series of animations displays (just via wire frame outline) where we are headed in terms of Recommendation. Users initially see a large single search box (like Google) and below it can see a reference to recommendations. After that we see a personalised home page with a basic dashboard offering recommendations much like Amazon. These include personal, curated and more general recommendations and then lead into social recommendations that may be curated with the context of the current record.

Recommendation

Generic: Related content

• keyword analysis
• machine learning (eg. OpenCalais)
• metadata

Targeted: Personalised recommendations

• user browsing history / click tracking
• ratings / reviews (eg. "I liked this")
• search history
• subject enrolment information

All users see the Generic recommendations and they are using the three elements outlined above. Most libraries are already working in this area at some basic level. The Targeted personalised recommendations are not so common. W e are looking at leveraging vast amounts of user data that we have to deliver targeted recommendations for particular users.

Recommendation

Open-source machine learning platform Easy to deploy Powerful Scalable

• distributed, parallel computing

Apache Mahout

A software platform currently available to simplify some of the complicated aspects of recommendation algorithms is Apache Mahout. It is currently used by Netflix. As it says above it is open-sourced, easy to deploy and use and offers powerful scaling and distribution capabilities. Scalability is key as data sources can grow and be combined from different sources.

Integration / Open

Javascript / jQuery RESTful services JSON / JSONP OAI-PMH

The other side of Discovery is to have the information machine

• discoverable. These are some examples of popular web technologies that

facilitate such machine discovery.

This series of animations just shows a concept for visual online browsing that was proposed by our 2012 Artist-in-Residence, Chris Gaul. It starts by visualising a Dewey arrangement of the entire collection arranged like a colourful DNA strip. Users could select a top-level subject and browse down through increasingly specific areas of interest to an individual item record. P .S. : When we presented at THETA I thought this was very much still at conceptual level and had not been further developed. I knew it was helping us to reshape our thinking and in taking our first steps towards Discovery and away from the more traditional Search methodologies. Only a couple

• f weeks later, our ITS team quietly released a working version of this

visual browsing concept on our catalogue it allows you to browse visually by author, title and subject and see where the results sit across the whole spectrum of our collection. Y

• u can also just browse down and into the

catalogue by selecting a colour (or top level subject area). MMB

What’s Possible

Library Collection

000 100 200 300 400 500 600 700 800 900 924,903 Items

This series of animations just shows a concept for visual online browsing that was proposed by our 2012 Artist-in-Residence, Chris Gaul. It starts by visualising a Dewey arrangement of the entire collection arranged like a colourful DNA strip. Users could select a top-level subject and browse down through increasingly specific areas of interest to an individual item record. P .S. : When we presented at THETA I thought this was very much still at conceptual level and had not been further developed. I knew it was helping us to reshape our thinking and in taking our first steps towards Discovery and away from the more traditional Search methodologies. Only a couple

• f weeks later, our ITS team quietly released a working version of this

visual browsing concept on our catalogue it allows you to browse visually by author, title and subject and see where the results sit across the whole spectrum of our collection. Y

• u can also just browse down and into the

catalogue by selecting a colour (or top level subject area). MMB

What’s Possible

Library Collection

000 100 200 300 400 500 600 700 800 900 924,903 Items

This series of animations just shows a concept for visual online browsing that was proposed by our 2012 Artist-in-Residence, Chris Gaul. It starts by visualising a Dewey arrangement of the entire collection arranged like a colourful DNA strip. Users could select a top-level subject and browse down through increasingly specific areas of interest to an individual item record. P .S. : When we presented at THETA I thought this was very much still at conceptual level and had not been further developed. I knew it was helping us to reshape our thinking and in taking our first steps towards Discovery and away from the more traditional Search methodologies. Only a couple

• f weeks later, our ITS team quietly released a working version of this

visual browsing concept on our catalogue it allows you to browse visually by author, title and subject and see where the results sit across the whole spectrum of our collection. Y

• u can also just browse down and into the

catalogue by selecting a colour (or top level subject area). MMB

What’s Possible

Languages

400 410 420 430 440 450 460 470 480 490 12,043 Items

This series of animations just shows a concept for visual online browsing that was proposed by our 2012 Artist-in-Residence, Chris Gaul. It starts by visualising a Dewey arrangement of the entire collection arranged like a colourful DNA strip. Users could select a top-level subject and browse down through increasingly specific areas of interest to an individual item record. P .S. : When we presented at THETA I thought this was very much still at conceptual level and had not been further developed. I knew it was helping us to reshape our thinking and in taking our first steps towards Discovery and away from the more traditional Search methodologies. Only a couple

• f weeks later, our ITS team quietly released a working version of this

visual browsing concept on our catalogue it allows you to browse visually by author, title and subject and see where the results sit across the whole spectrum of our collection. Y

• u can also just browse down and into the

catalogue by selecting a colour (or top level subject area). MMB

What’s Possible

Languages

400 410 420 430 440 450 460 470 480 490 12,043 Items

This series of animations just shows a concept for visual online browsing that was proposed by our 2012 Artist-in-Residence, Chris Gaul. It starts by visualising a Dewey arrangement of the entire collection arranged like a colourful DNA strip. Users could select a top-level subject and browse down through increasingly specific areas of interest to an individual item record. P .S. : When we presented at THETA I thought this was very much still at conceptual level and had not been further developed. I knew it was helping us to reshape our thinking and in taking our first steps towards Discovery and away from the more traditional Search methodologies. Only a couple

• f weeks later, our ITS team quietly released a working version of this

visual browsing concept on our catalogue it allows you to browse visually by author, title and subject and see where the results sit across the whole spectrum of our collection. Y

• u can also just browse down and into the

catalogue by selecting a colour (or top level subject area). MMB

What’s Possible

Linguistics

1,918 Items 415 410 411 412 413 414 417 418 419 416

This series of animations just shows a concept for visual online browsing that was proposed by our 2012 Artist-in-Residence, Chris Gaul. It starts by visualising a Dewey arrangement of the entire collection arranged like a colourful DNA strip. Users could select a top-level subject and browse down through increasingly specific areas of interest to an individual item record. P .S. : When we presented at THETA I thought this was very much still at conceptual level and had not been further developed. I knew it was helping us to reshape our thinking and in taking our first steps towards Discovery and away from the more traditional Search methodologies. Only a couple

• f weeks later, our ITS team quietly released a working version of this

visual browsing concept on our catalogue it allows you to browse visually by author, title and subject and see where the results sit across the whole spectrum of our collection. Y

• u can also just browse down and into the

catalogue by selecting a colour (or top level subject area). MMB

Participation Privacy Concerns

Challenges

Two challenges remain:

• a. encouraging participation in our initiatives (because they fail without

participation), - so it has to be easy, attractive and deliver on promises; and

• b. addressing the privacy concerns that some people may have about the

collection and sharing of data about their habits in the Library. It is best taken care of at the outset rather than in response to user complaints or

• fears. W

e need to assure potential users that it is an opt-in service and their data is subject to protections against misuse.

THANK YOU

D E S I G N : T O M F E T H E R S C H R I S G A U L R E S E A R C H : J U L I E - A N N E M A R S H A L L B E L I N D A T I F F E N

UTS: LIBRARY