How I Spent My Summer - PowerPoint Presentation

Slide1

How I Spent My Summer

– or –

Oxford-Illinois Digital Libraries Placement Program

Summer

2015

Jennifer Westrick, MSLIS

University of Illinois, OIDLPPSlide2

The Initial Project

Project

3: Migration Workflow for Digital Collections.

The Bodleian has been digitizing collections and making them available online for more than twenty years. While the images created can still be useful and viable, the platforms to deliver the images are often difficult to maintain. This project will analyze existing digitization and publication workflows, and propose ways of making legacy content and collections available online through new platforms.

Written by Christine Madsen, Head of Digital

Programmes

, BDLSS

(Bodleian Digital Libraries Systems and Services)Slide3

Project Changed to Become an Exploration of the Current Situation

Before moving ahead on designing workflows, basic information was lacking:

- How many are we talking about?

- What platforms are they built on?

- Does anyone still use them?

Digital collections constantly changing; workflow still

neededSlide4

Deliverables

:

- Spreadsheet

- Detailed walk-through/ workflow of

steps taken

(

what

worked

and what didn’t

)

Slide5

Retiring a Digital Library:

Considerations for Legacy

Digital

Collections

Jennifer Westrick, MSLIS

University of Illinois, OIDLPPSlide6

d

igital.bodleian

was

i

ntroduced in July 2015Slide7

But what

about these?Slide8

But what about these?Slide9

But what

about these?Slide10

But

what about these?Slide11

But w

hat about these?Slide12

But

what about these?Slide13

But w

hat about these?Slide14

But w

hat about these?Slide15

These legacy collections

Are not an issue at smaller and/or less progressive institutions

Grew

organically; often part of a funded project with specified time frame

May be currently used

May have important sites that reference/cite the URL

Are

(?) in

digital.bodleianSlide16

These legacy collections

Are not an issue at smaller and/or less progressive institutions

Grew

organically; often part of a funded project with specified time frame

May be currently used

May have important sites that reference/cite the URL

Are

(?) in

digital.bodleian

Goal: easy for users, and less work for IT

Slide17

The Process: Five

Steps

Define: what is a legacy collection?

Identify: The goal is a comprehensive list of possible legacy collections

Collect data

- Basic information about

site

- Patron usage (Google Analytics)

- Scholarly usage (Google Scholar and Webometrics)

- how many outside URLs link to that digital collection

- source of that link

- Technical

Data

Assess the collections

Next steps for the collectionsSlide18

Step One:

Define

Goal

:

what criteria must a collection meet to be considered legacy?

A legacy collection:

- its content is now part of a newer collection

- its technology managed by Bodleian IT staff

- some judgement calls, i.e. blogs, exhibits

- don’t forget collections within collectionsSlide19

Step Two: Identify

Goal

: a

list of legacy

collections

Sources:

- spreadsheet from Christine Madsen

- spreadsheet from Michael

Popham

- quick Google search

- list of sites from IT?Slide20

Step Three: Collect

Data

Goal: Numbers in spreadsheet (color-coded!)

Four areas of analysis

- Basic information about site

- Patron usage (Google Analytics)

- Scholarly usage (Google Scholar and Webometrics)

- how many outside URLs link to that digital collection

- source of that link

- Technical DataSlide21

Basic informationSlide22

Patron U

sage: Google Analytics

- Google Analytics currently tracking usage on 23 of the legacy sites

- website statistics include a 30-day summary of

- sessions

-

unique

users

-

pageviews

- pages

per session

-

avg

session duration

-

%

of users

who were

new

- screenshots saved for future analysis

Slide23

2,259 sessions

1,776 users

8,601

pageviews

3.81 pages per session

03:06 avg. session duration

72% new sessions

Over

the past 30 days,

ballads.bodleian.ox.ac.uk hadSlide24

Google: 1,460

Direct logins: 376

Referral: 171

(harkavagrant.com)

Bing: 39

Yahoo: 13

Sources of

the 2,259 sessions

:Slide25
Slide26
Slide27
Slide28

Other means to analyse patron usage?

-

Software that require purchase

- Google keyword search:

- hard to do accurately - search results vary widely depending on keyword used

- Google assumptions - tailoring results

- Alexa or other web-ranking sites: can’t get them to take entire URL

“

arshama.bodleian.ox.ac.uk” searches on “

ox.ac.uk

”Slide29

URL Analysis: Google Scholar and Webometrics

Assumption:

Greater number of documents that include a reference, citation

or link to our URL =

greater significance in the scholarly world (and less inclination to change the URL)Slide30

Google Scholar search:

G

oal

:

citation information – how many times was the legacy collection’s URL cited?

Go

into Google Scholar, search on

URL

- # mentions (supplied by Google Scholar, at the top of the page)

-

# of items that cite the URL (count each item once

)

- # of citations (how many total citations from the above items) Slide31

3 items have cited this URL

81 total results

..for a total of 44 citations

↗

↗

↗Slide32
Slide33

Google

Scholar can be:

- a source of detailed information about each URL that cited the collection

(started compiling; too much data)

- searched many ways – i.e. root vs entire URL

- managed; many academics and institutions monitor their inclusions Slide34

URL

Analysis with

Webometrics

Different

from Google Scholar because

it returns URL usage on websites rather than publications, and no focus on academia.

Best thing: List

of

URLs

Mike

Thelwall’s

Statistical

Cybermetrics

Research Group

Webometrics

Analyst

2.0, http

://lexiurl.wlv.ac.uk/index.html Slide35

Webometrics screen shot

Search this list of URLs for wikip, .edu, .ac.uk; record totalsSlide36
Slide37

Technical information

Goal: collect specifics about each site to assess the difficulty of future actions.

Problem: Information is not easily accessible.Slide38

Technical information

Goal

: collect specifics about each site to assess the difficulty of future actions.

Problem: Information is not easily accessible.

Also collect information on metadata, if it existsSlide39

Compilation

Approximately 150 rows of potential legacy sites, and 45 columns of criteria Slide40

Step Four: Assess

Helpful

to mark the collection if it is

- Active

- Dead

link

- Duplicate

or near-duplicate

-Not

ours:

An exhibit

A different department or different university

entirely

Is

it in

digital.bodleian

?Slide41

Step Five: Potential Next Steps

Goals

: Maintain user access and

lighten the

workload on IT

- delete the website - redirect: automatically or with a click

- pros: seamless to user

- con: site still needs to be maintained

- move all legacy collections to a single server

- make the pages static and put all on a single server

- Archive-It - or other outside web archiving serviceSlide42

Potential Next Steps – big

picture

- Create

guidelines for future digital

collections

- specify

supported platforms

- metadata specifics

- regularly scheduled maintenance and/or checks

- Updated Digital Collections Management Policy

- Much depends on staff/time/budgetSlide43

Retiring

a Digital Library:

Considerations for Legacy Digital

Collections

Deliverables:

-

a detailed walk-through of steps taken (what worked and what didn’t)

- a spreadsheet

with lots of data (color-coded!)

Jennifer

Westrick, MSLIS

University of Illinois,

OIDLPP

Summer

2015