Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
AnalyzingDigitalCollectionsEntrances:WhatGetsUsedandWhyItMatters
ParomitaBiswasandJoelMarchesoni
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016
19
ABSTRACT
ThispaperanalyzesusagedatafromHunterLibrary’sdigitalcollectionsusingGoogleAnalyticsforaperiodoftwenty-sevenmonthsfromOctober2013throughDecember2015.Theauthorsconsiderthisdataanalysistobeimportantforidentifyingcollectionsthatreceivethelargestnumberofvisits.Wearguethisdataevaluationisimportantintermsofbetterinformingdecisionsforbuildingdigitalcollectionsthatwillserveuserneeds.TheauthorsalsostudythebenefitsofharvestingtositessuchastheDigitalPublicLibraryofAmerica,andtheybelievethispaperwillcontributetotheliteratureonGoogleAnalyticsanditsusebylibraries.INTRODUCTION
HunterLibraryatWesternCarolinaUniversity(WCU)hasfourteendigitalcollectionshostedinCONTENTdm—adigitalcollectionmanagementsystemfromOCLC.Userscanenterthecollectionsinvariousways—throughtheLibrary’sCONTENTdmlandingpages,1searchengines,orsitessuchastheDigitalPublicLibraryofAmerica(DPLA)whereallthecollectionsareharvested.2SinceOctober2013,theLibraryhascollectedusagedatafromitscollections’websitesandfromDPLAreferralsviaGoogleAnalytics.Thispaperanalyzesthisusagedatacoveringaperiodofapproximatelytwenty-sevenmonthsfromOctober2013throughDecember2015.Theauthorsconsiderthisdataanalysisimportantforidentifyingcollectionsreceivingthelargestnumberofvisits,includingvisitsthroughharvestingsitessuchastheDPLA.Theauthorsarguethatsuchdataevaluationisimportantbecauseitcanbetterinformdecisionstakentobuildcollectionsthatwillattractusersandservetheirneeds.Additionally,thisanalysisofusagedatageneratedfromharvestingsitessuchastheDPLAdemonstratestheusefulnessofharvestinginincreasingdigitalcollections’usage.Lastly,thispapercontributestothebroaderliteratureonGoogleAnalyticsanditsusebylibrariesindataanalysis.
LITERATUREREVIEW
UsingGoogleAnalyticstostudyusageofelectronicresourcesiscommon;aconsiderableamountofmaterialexistsdescribingtheuseofGoogleAnalyticsinmarketingandbusinessfields.3
ParomitaBiswas([email protected])isMetadataLibrarianandJoelMarchesoni([email protected])isTechnologySupportAnalyst,HunterLibrary,WesternCarolinaUniversity,Cullowhee,NorthCarolina.
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
20
However,thepublishedliteratureofferslittleabouttheuseofthissoftwareforstudyingusageofcollectionsconsistingofuniquematerialsdigitizedandplacedonlinebylibrariesandculturalheritageorganizations.Forexample,BettyhaswrittenaboutusingGoogleAnalyticstotrackstatisticsforuserinteractionwithlibrarian-createddigitalmediasuchasquizzesandvideotutorials.4FangdiscussesusingGoogleAnalyticstotrackthebehaviorofuserswhovisitedtheRutgers-NewarkLawLibrarywebsite.5Fanglookedatthenumberofvisitors,whatandhowmanypagestheyvisited,howlongtheystayedoneachpage,wheretheywerecomingfrom,andwhichsearchengineorwebsitehadreferredthemtothelibrary’swebsite.Findingswereevaluatedandusedtomakeimprovementstothelibrary’swebsite.Forexample,FangmentionsusingGoogleAnalyticsdatafortrackingthepercentageofnewandreturningvisitorsbeforeandafterthewebsiteredesign.
Amongarticlesthatdiscussusingwebanalyticstolearnhowusersaccessdigitalcollections,mosthavefocusedonacomparisonbetweenthird-partyplatforms,onlinesearchengines,andthetraditionallibrarycatalogtofindpreferredmodesofaccessandwhetherresultscallforashiftinhowlibrariessharetheirdigitalcollections.Forexample,intheirarticleontheimpactofsocialmediaplatformssuchasHistoryPinandPinterestonthediscoveryandaccessofdigitalcollections,BaggettandGibbsuseGoogleAnalyticsfortrackingusageofdigitalobjectsonthelibrary’swebsiteaswellstatisticscollectedfromHistoryPin’sandPinterest’sfirst-partyanalyticstools.6TheauthorsconcludethatwhileneitherHistoryPinnorPinterestdriveusersbacktothelibrary’swebsite,theyhelpinthediscoveryofdigitalcollectionsandcanenhanceuseraccesstolibrarycollections.SchlosserandStampercomparetheeffectsonusageofacollectionhousedinaninstitutionalrepositoryandrepostedonFlickr.7Whetherhousingacollectiononathird-partysitehadanadverseeffectonattractingtraffictothelibrary’swebsitewasnotasimportantasensuringusersaccessedthecollectionsomewhere.Likewise,O’EnglishdemonstrateshowdatafromwebanalyticswereusedtocompareaccesstoarchivalmaterialsviaonlinesearchenginesasopposedtolibrarycatalogsusingMARCrecordsfordescriptions.8O’Englishargueslibrarypracticesshouldchangeaccordinglytopromotepatronaccessanduse.Ladd’sarticleontheaccessanduseofadigitalpostcardcollectionfromMiamiUniversityusesstatisticsfromGoogleAnalytics,CONTENTdm,andFlickroveraperiodofoneyear.9Ladd’sfindingsrevealthatfewuserscametothemaindigitalcollectionswebsitetosearchandbrowse;instead,mostarrivedviaexternalsourcessuchassearchenginesandsocialmediasites.Theresultingincreaseinviewsmakesitimperative,Laddasserts,thatregularupdatesbothinCONTENTdmandFlickrareimportantforpromotingaccessanduseofthepostcards.
ArticlesonusingGoogleAnalyticsfortrackingdigitalcollectionusagehaveexploredtrackingthegeographicbaseofusers.Forexample,HeroldusesGoogleAnalyticstodemonstrateusageofadigitalarchivalcollectionbyusersatinstitutional,national,andinternationallevels.10HeroldlooksatservertransactionlogsmaintainedinGoogleAnalytics,on-andoff-campussearchingcounts,userlocations,andrepeatvisitorstothearchivalimagesrepresentingculturalheritagematerialsrelatedtoOrangAslipeoplesandculturesofMalaysia.Sheusesthesedatatoascertain
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016 21
thenumberofusersbygeographicregionanddeterminethat,whilemostvisitorscamefromtheUnitedStates,Malaysiarankedsecond.Thedatasupported,accordingtoHerold,thatthisparticulardigitalcollectionwasabletoreachanothertargetaudience:usersfromMalaysia.Herold’sfindingsindicatethatdigitizationofuniquematerialsmakesthemavailabletoaworldwideaudience.
WhetherharvestinghasincreasedusageofdigitalcollectionsavailableviaDPLAoritshubshasreceivedlimitedexplorationintheliterature.Mostwritingsonharvestingdigitalcollectionshavefocusedmoreonthetechnicalaspectsoftheprocess,liketheDPLA’singestionmethod,thequalityandscalabilityofmetadataremediationandenhancement,11andlargemetadataencoding.12Forexample,GregoryandWilliamswriteabouttheNorthCarolinaDigitalHeritageCenterasoneoftheservicehubsoftheDPLA.TheservicehubsarecentersthataggregatedigitalcollectionmetadataprovidedbyinstitutionsforharvestingbytheDPLA.Theauthorsdiscussmetadatarequirements,softwarereview,andestablishmentofworkflowforsendinglargemetadatafeedstotheDPLA.13Boyd,Gilbert,andVinson,intheirarticleontheSouthCarolinaDigitalLibrary(SCDL),anotherservicehubforDPLA,describetheplanningbehindsettinguptheSCDL,itsmanagement,andthetechnologyinvolvedinmetadataharvesting.14FreelandandMoulaisondiscusstheMissourihubasamodelfor“institutionswithsimilarcollectivegoalsforexposingandenrichingtheirdatathroughtheDPLA.”15Accordingtothem,byharvestingtheirmetadatatotheDPLA,institutionsareabletosharetheirdigitalcollectionswiththebroaderpublic.Additionally,institutionsthatharvestmetadatatotheDPLAgetvalue-addedserviceslikegeocodingoflocation-basedmetadataandexpressionofcontributedmetadataaslinkeddata.
DataCollectionParameters
HunterLibrarydigitalcollectionsusagedataincludedinformationonitemviews16andreferrals17foreachofthecollectionsincludingDPLAreferrals.Theauthorsalsoconsideredkeywordsearchterms18acrossallreferrals,andwithinCONTENTdmspecifically,thatbroughtuserstotheLibrary’scollections.Theauthorsconsideredthemostfrequentlyoccurringkeywordstoberepresentingthesubjectsofcollectionsthatweremostused.RepeatvisitorstotheLibrary’sdigitalcollections’websitewerealsotracked.Finally,sessions19weretracedbythegeographicarea20oftheusers.
HunterLibrary’scollectionsvaryinsize.TheLibrary’slargestandoneoftheoldestcollections,CraftRevival[Note:collectionsaresetinromanandcapitalized]showcasesdocuments,photographs,andcraftobjectshousedinHunterLibraryandsmallerregionalinstitutions.Thecollection’sitemsrepresentthelatenineteenthandearlytwentiethcentury(1890s–1940s)CraftRevivalmovementinWesternNorthCarolina,whichwascharacterizedbyarenewedinterestinhandmadeobjects,includingCherokeeartsandcrafts.TheCraftRevivalcollectionbeganin2005andincludes1,982items.Thesecondlargestcollection,GreatSmokyMountains,whichhighlightseffortsthatwentintotheestablishmentoftheparkandincludesphotographsonthelandscapeandfloraandfaunainthepark,beganin2012andconsistsof1,829items.Notalldigital
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
22
collectionswereharvestedtotheDPLAatthesametime.WhilesomeoldercollectionswereharvestedtotheDPLAin2013,smaller,institution-specificcollectionsstartedlaterwerealsoharvestedlater.ForexampleWCU—OralHistories,acollectionofinterviewscollectedbystudentsofoneofWCU’shistoryclassesdocumentingthehistoryandcultureofWesternNorthCarolinaandthelivesofWCUathletesorartists’likeJosephinaNiggliwhotaughtdramaatWCU;HighlightsfromWCU,acollectionofuniqueitemsfromWCU’sMountainHeritageCenterandotherdepartmentsoncampus,includinglettersfromtheLibrary’sSpecialCollectionstranscribedbyWCU’sEnglishdepartmentstudents;andWCU—FineArtMuseum,showcasingartworkfromtheuniversity’sFineArtMuseum,wereharvestedtotheDPLAin2015.Asthesesmallercollectionswerestartedlater,theirtotalitemviewsandreferralcountswouldlikelybelessthansomeoftheLibrary’soldercollections;however,thesenewercollectionswereincludedastheymightprovidevaluabledataregardingharvestingreferralsandreturningvisitors.Table1showstheyearsthecollectionswerestarted,thenumberofitemsincludedineachcollection,andtheyeartheywereharvestedtotheDPLA.
CollectionName StartYearCollectionSize
(NumberofItems) HarvestedSince
CherokeeTraditions 2011 332 2013CivilWar 2011 68 2013CraftRevival 2005 1,982 2013GreatSmokyMountains
2013 1,829 2013
HighlightsfromWCU 2015 39 2015HoraceKephart 2005 552 2013PicturingAppalachia 2012 972 2013StoriesofMountainFolk
2012 374 2013
TravelWesternNorthCarolina
2011 160 2013
WCU—FineArtMuseum
2015 87 2015
WCU—Herbarium 2013 91 2013WCU—MakingMemories
2012 408 2013
WCU—OralHistories 2015 67 2015WesternNorthCarolinaRegionalMaps
2015 37 2015
Table1.Collectionsbyyear
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016 23
CollectingDataUsingGoogleAnalytics
TheLibraryhashadGoogleAnalyticssetupononlineexhibits—websitesoutsideofCONTENTdmthatprovideadditionalinsightintothecollection—since2008andbeganusingGoogleAnalyticstotrackitsCONTENTdmmaterialswiththe6.1.2releaseinOctober2013.CONTENTdmversion6.4introducedaconfigurationfieldthatallowedtheauthorstoenteraGoogleAnalyticsIDandautomaticallygeneratethetrackingcodeinpagestosimplifythesetup.Followingthatsoftwareupdate,OCLCmadeGoogleAnalyticsthedefaultdataloggingmechanism.
TheLibrarysetupGoogleAnalyticssuchthatonlineexhibitsaretrackedtogetherwiththeirCONTENTdmcollections.ThisisaccomplishedbyusingcustomtrackingonallwebpagesandacustomscriptinCONTENTdm.ThisallowstheLibrarytolinkitsCONTENTdmandwcu.edudomainswithinGoogleAnalyticssothatsessionscanbeviewedacrossallonlinedigitalcollections.
DatawerecollectedfromGoogleAnalyticsusingseveraltools.GoogleprovidesanonlinetoolcalledQueryExplorer(https://ga-dev-tools.appspot.com/query-explorer/)thatcancreateandexecutecustomsearchesagainstGoogleAnalytics.Thisapplicationwasusedtocraftthequeries.MicrosoftExcelwasprimarilyusedtodownloaddata,usingthecustompluginResttoExcelLibrary(http://ramblings.mcpher.com/Home/excelquirks/json/rest)toparseinformationfromGoogleAnalyticsintoworksheets.TheExceladd-onworkswell,butrequiresknowledgeofMicrosoftVisualBasicforApplications(VBA)programmingtouseeffectively.Thislimitationpromptedtheauthorstolookforasimplerwayofretrievingdata.TheauthorsfoundOpenRefine(https://github.com/OpenRefine/OpenRefine)tocollect,sort,andfilterdata,withExcelusedforresultsanalysis.OnceinExcel,formulaswereusedtominedataforspecifictargets.
RESULTSANALYSIS
ThedatacollectedusingGoogleAnalyticsspannedaperiodofapproximatelytwenty-sevenmonths,fromOctober2013throughDecember2015.Table1andgraph1showeachcollection’sitemviews,itemreferrals,andsize(numberofitemsinthecollection).Thesenumberswerecalculatedforeachcollectionasapercentageoftotalitemviews,totalitemsreferrals,andtotalnumberofitemsforallcollectionstogether.Intable2,thetopfivecollectionsintermsofitemsviewsandreferralsarehighlighted.Graph1,agraphicalrepresentationoftable2,displaysmorestarklythedifferencesbetweencollectionsintermsofviewsandreferrals.
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
24
CollectionName
ItemViewsasPercentageofTotal
Views
ItemReferralsasPercentageofTotalReferrals
NumberofItemsasPercentageofTotalItems
forallCollectionsCherokeeTraditions 6.38 6.12 4.74CivilWar 1.89 0.88 0.97CraftRevival 41.35 52.39 28.32GreatSmokyMountains 7.50 6.34 26.14HighlightsfromWCU 0.23 0.08 0.56HoraceKephart 11.67 7.62 7.89PicturingAppalachia 10.03 9.99 13.89StoriesofMountainFolk 3.51 2.45 5.344TravelWesternNorthCarolina 7.87 9.57 2.29
WCU—FineArtMuseum 0.19 0.08 1.24WCU—Herbarium 0.71 0.45 1.30WCU—MakingMemories 7.13 2.64 5.83WCU—OralHistories 0.80 1.08 0.96WesternNorthCarolinaRegionalMaps 0.26 0.11 0.53
Total 100.00 100.00 100.00Table2.Collectionsbypercentage
Graph1.Collectionsbypercentage
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016 25
Asdemonstratedintheprecedingtableandgraph,CraftRevival,oneoftheLibrary’soldestandlargestcollections,contributesmorethan28percentofalldigitalcollections’itemsandgarnerscloseto42percentofallitemviewsand53percentofallitemreferrals.GreatSmokyMountains,thesecondlargestcollection,contributesalittlemorethan26percentofitemsbutreceivesonlyabout8percentofallitemviewsand7percentofallreferrals.TheHoraceKephartcollection,focusingonthelifeandworksofHoraceKephart—author,librarian,andoutdoorsmanwhomadethemountainsofWesternNorthCarolinahishomelaterinlife—istheLibrary’sfourthlargestcollection.Itreceivesalmost12percentofallitemviewsandabout8percentofallitemreferrals.PicturingAppalachia,thethirdlargestcollection—consistingofphotographsshowcasingthehistory,culture,andnaturallandscapeofSouthernAppalachiaintheWesternNorthCarolinaregion—makesup14percentofitemsandreceivesapproximately10percentofallreferralsandviews.TravelWesternNorthCarolina—visualjourneysofWesternNorthCarolinacommunitiesthroughthreegenerations—contributesfewerthan3percentofitemsbutscoreshighonbothitemsviewsandreferrals.WCU—MakingMemories,whichhighlightsthepeople,buildings,andeventsfromWCU’shistory,andStoriesofMountainFolk(SOMF),whichisacollectionofradioprogramsfromWesternNorthCarolinanon-profitCatchtheSpiritofAppalachiaandarchivedatHunterLibrary,arecollectionsthataresimilarinsize—receivingfewerthan3percentofallitemreferrals.However,WCU—MakingMemoriesreceivesamorethan7percentofallitemviewscomparedtoSOMF’salmost4percent.
ThesefindingsarenotsurprisingastheMakingMemoriescollectiondocumentsWesternCarolinaUniversity’shistoryandmayreceivemanyviewsfromwithintheinstitution.Overall,however,theCraftRevivalcollectioncanbeconsideredtheLibrary’smostpopularcollection.TheHoraceKephartcollectionappearstobethesecondmostpopularcollection.And,notsurprisingly,CherokeeTraditions,acollectionofartobjects,photographs,andrecordingssimilarincontenttotheCraftRevivalintermsofitsfocusonCherokeecultureandhistory,isquitepopularandreceivesmoreitemreferralsthanbothWCU—MakingMemoriesandSOMFandmoreitemviewsthanSOMF(table2).
AnanalysisofkeywordsearcheswithinCONTENTdmandkeywordsearchesacrossallreferralsourcesreiteratesthesefindings.Aspartoftheanalysis,datacollectedforthistwenty-seven-monthperiodforthetopkeywordsearcheswithinCONTENTdmandthetopkeywordsearchescountingallreferralswasrecordedinanExcelspreadsheetandthenuploadedtoOpenRefine.OpenRefineallowstextandnumericdatatobesortedbyname(alphabetical)andcount(highesttolowestoccurring).OncetheExcelspreadsheetwasuploadedtoOpenRefine,keywordsweresortednumericallyandclustered.OpenRefinehasa“cluster”functiontobringtogethertextthathasthesamemeaningbutdiffersbyspellingorcapitalization(forexample,“CHEROKEE,”“cherokee,”“cheroke”)orbyorder(forexample,“JaneSmith,”“Smith,Jane”).Theclusteringfunctionprovidesacountofthenumberoftimesakeywordwasusedregardlessofexactspelling.Afteridentifyingkeywordsbelongingtoacluster(forexample,aclusteroftheword“Cherokee”spelleddifferently),thedifferentlyspelledororganizedkeywordsineachclusterweremergedin
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
26
OpenRefinewiththeirmostaccuratecounterparts.Finally,itshouldbenotedthatkeywordsincluding“!”and“+”symbolsweremostlikelygeneratedfromeitherusingmultiplesearchtermswithinCONTENTdm’sadvancedsearchorfromcuratedsearchlinksmaintainedonsomeofouronlineexhibitwebsites.Theselinkstakeuserstocommonlyusedresultsetswithinthecollection.Tables3and4providealistingofthetenmostfrequentlysearchedkeywordswithinCONTENTdmacrossallreferralsandnamesofcollectionsthataremostrelevanttothesesearches.
Keywords OccurrenceCount RelevantCollection(s)
Cherokee 187 CraftRevival;CherokeeTraditions
CherokeeLanguage 107 CraftRevival;CherokeeTraditions
SouthernHighlandCraftGuild
98 CraftRevival
basket!object 96 CraftRevival;CherokeeTraditions
Indianmasks—AppalachianRegion,Southern
83 CraftRevival;CherokeeTraditions
Basket!photographpostcard 82 CraftRevival;CherokeeTraditions
W.M.ClineCompany 78 PicturingAppalachia;CraftRevival
Cherokee+Indian!photograph
72 CraftRevival;CherokeeTraditions
Wood-carving—AppalachianRegion,Southern
70 CraftRevival
Indianwood-carving—AppalachianRegion,Southern
69 CraftRevival
Table3.TopkeywordssearcheswithinCONTENTdm
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016 27
Keywords NumberofSessions RelevantCollection(s)cherokeetraditions 442 CraftRevival;Cherokee
Traditionshoracekephart 185 HoraceKephart;Great
SmokyMountains;PicturingAppalachia
cherokeepottery 55 CraftRevival;CherokeeTraditions
kephartknife 50 HoraceKephartamandaswimmer 37 CraftRevival;Cherokee
Traditionsappalachianpeople 36 CraftRevival;Cherokee
Traditions;GreatSmokyMountains;WCU—Oral
Historiescherokeeindianpottery 36 CraftRevival;Cherokee
Traditionscherokeebaskets 34 CraftRevival;Cherokee
Traditionsweavingpatterns 33 CraftRevival;Cherokee
Traditionsbasketweaving 26 CraftRevival;Cherokee
Traditions
Table4.Topkeywordsearchesacrossallreferrals
Tables3and4showthattopsearchesrelatetoartsandcraftsfromtheWesternNorthCarolinaregion(“baskets,”“Indianmasks,”“Indianwoodcarving,”“Cherokeepottery”),artists(“amandaswimmer”),ortopicsrelatingtoCherokeeculture(“cherokee,”“cherokeelanguage”).SearchesrelatingtotheHoraceKephartcollection(“horacekephart,”“kephartknife”)arealsopopular,explainingthefactthattheKephartcollection,whichaccountsforfewerthan8percentoftheLibrary’sdigitalcollections’itemsscoreshighlyintermsofitemviews(second)andreferrals(fourth).
ThepopularityoftopicsrelatedtoWesternNorthCarolinaisreiteratedinthegeographicbaseoftheusers.Graph2showsNorthCarolinaaccountsformostofthesearches,withcitiesinWesternNorthCarolina(Asheville,Franklin,Cherokee,Waynesville)accountingformorethan40percentofsessions.
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
28
Graph2.Citiesbysessioncount
ThemajorityofitemreferralscomefromsearchenginessuchasGoogle,Bing,andYahoo!Graph3showsthepercentageofitemreferralsfromtheseexternalsearches.21However,theDPLAalsogeneratesafairamountofincomingtraffictothecollections.Forexample,whileallcollectionsgetreferralsfromtheDPLA,harvestingtotheDPLAisparticularlyusefulforsmallercollectionssuchasHighlightsfromWCU,WCU—FineArtMuseum,andCivilWarCollection.Eachofthesecollectionsgets17percentofreferralsfromtheDPLA,makingDPLAthelargestreferralsourcefollowingthesearchenginesfortheHighlightsandFineArtMuseumcollections.Graph4showsreferralseachcollectionreceivesviatheDPLAasapercentageoftotalreferrals.ThisindicatestheusefulnessofharvestingtotheDPLA.AtrendseemsalsotoshowthereisanincreaseintotalreferralsfromDPLApermonththelongeritemsareinDPLA(graph5).
Graph3.Percentageofsearchengineitemreferrals(Google,Bing,andYahoo!)
367319
171146 144 135 122 109 105 98
44%
29%
47% 44%
75%
43% 57%
11% 23%
75% 74%
38% 33%
6% 22%
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016 29
Graph4.PercentageofDPLAitemreferrals
Graph5.IncreaseinDPLAreferralsovertime
Lastly,newandreturningvisitorstothecollectionsweretrackedasamarkerofuserinterestinparticularcollections.Graph6showsdatacollectedfornewandreturningvisitorscalculatedasaproportionofthetotalnumberofvisitsforeachcollection.SomesmallercollectionslikeHighlightsfromWCU,WNCRegionalMaps,WCU—FineArtMuseum,andWCU—OralHistoriesscorehighlyintermsofattractingreturnvisitors(graph6).
6%
17%
3%
12%
17%
4%
11%
6% 3%
17%
3% 4% 5% 0%
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
30
Graph6.Newandreturningvisitors
DISCUSSION
TheaimbehindgatheringdatawastostudyusageofHunterLibrary’sdigitalcollectionsandexaminetheusefulnessofharvestinginpromotinguse.Althoughusagedatalogswereunabletoshedmuchlightontheactualusefulnessofthecollectionstousers,thelogsprovidedinformationonvolumeofuse,whatmaterialswereaccessed,andwhereuserswerelocated.Analysisofthetransactionlogsindicatesthatwhileallcollectionslikelybenefittedfromharvesting,CraftRevival,CherokeeTraditions,andHoraceKephart(collectionsfocusingonthecultureandhistoryofwesternNorthCarolina)werethemostheavilyusedandmostvisitorscamefromthestateofNorthCarolinaandfromtheregioninparticular.SearchtermsinthetransactionlogsalsoindicatedastronginterestinitemsrelatedtoCherokeecultureandHoraceKephart.AsHerold,whotracedthesecondlargestgroupofusersoftheOrangAslidigitalimagearchivetoMalaysianotes,thegeographicbaseofacollection’suserscanbeindicativeofthepopularityofasubjectarea.22Likewise,Matusiakassertsthatusers’commentscanbeindicativeoftherelevanceofcollectionstousers’needsandprovidedirectionforthefuturedevelopmentofdigitalcollections.23
AsneithertheCraftRevival,CherokeeTraditions,norHoraceKephartcollectionincludesitemsthatrelatespecificallytotheuniversity’shistory—unlikeotherinstitution-specificcollectionsmentionedearlier—itispossiblecollectionusersmaybemorerepresentativeofthelargerpublicthantheuniversity.Thesefindingspointtotheneedforquestioningidentificationofanacademic
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016 31
library’suserbaseasmainlystudentsandfacultyoftheinstitutionandwhetherlibrariansshouldgivegreaterconsiderationtotheneedsofawideraudience.24Datasupportingtheexistenceofthisuserbase,whosetrueimportorpreferencesmightnotbecapturedinsurveysandquestionnaires,canserveasavaluablesourceofinformationforindividualsresponsibleforbuildingdigitalcollections.
InaninformalsurveyofHunterLibraryfacultycarriedoutbyHunterLibrary’sDigitalInitiativesUnitinSeptemberof2014,respondentsconsideredcollectionssuchasCraftRevivaltobemoreusefultousersexternaltotheuniversity.WhilethesurveycouldalludetothenatureoftheuserbaseofacollectionlikeCraftRevival,itunderstandablycouldnotcapturethescaleoftheitemviewsandreferralsgarneredbythiscollectionaswellasausagedataanalysiscould.Ontheotherhand,analysisofusagedata,asdemonstratedinthispaper,indicatedthatcertaincollections—HighlightsfromWCU,WCU—FineArtMuseum,andWCU—OralHistories—possiblyservedanicheaudience.Thesesmallerandmorerecentlyestablishedcollectionsconsistingofuniversity-createdmaterialsattractedmorereturningvisitors(seegraph6).Thesereturningvisitorswerelikelyinternaluserswhosevisitsindicated,asFangpointsout,aloyaltytothesecollections.25
Inthepaper“AFrameworkofGuidanceforBuildingGoodDigitalCollections,”authoredbytheNationalInformationStandardsOrganizationFrameworkAdvisoryGroup,theauthorspointoutthatwhiletherearenoabsoluterulesforcreatingqualitydigitalcollections,agoodcollectionshouldincludedatapertainingtousage.26Theauthorspointtomultipleassessmentmatrixesincludingusingacombinationofobservations,surveys,experiments,andtransactionloganalyses.AstheWCUdigitalcollectionsfindingsdemonstrate,acarefulanalysisofthepopularityofcollectionscanindicatetheneedforbalancingquantitativedatawithmorequalitativesurveyandinterviewdata.Thesefindingsalsoindicatethatusagedataanalysiscanbeveryvaluableinidentifyingtheextentofcollectionusagebyvisitorswhomaynothavesignificantsurveyrepresentation.Resultsfromthesmall(fewerthantenrespondents)WCUsurveyindicatethatsomerespondentsquestiontheinstitutionalusefulnessofcollectionssuchasCraftRevival.Theseresultsshowtheimportanceoftakingmultiplefactorsintoaccountwhenassessinguserneedsandinterestsindigitalcollections.
CONCLUSION
Theauthorsfeelfutureprojectsmightstemfromthisdataanalysis.Forexample,localsubjectfieldsbasedonthehighestrecurringkeywordsthatwereminedfromthetransactionlogscanbeaddedforallofHunterLibrary’sdigitalcollections.Usagestatisticsatalaterperiodcouldbeevaluatedtostudyifadditionofusergeneratedkeywordsincreaseduseofanycollection.AsMatusiakpointsoutinherarticleontheusefulnessofuser-centeredindexingindigitalimagecollections,socialtagging—despiteitslackofsynonymcontrolormisuseofthesingularandplural—isapowerfulformofindexingbecauseof“closeconnectionwithusersandtheirlanguage,”asopposedtotraditionalindexing.27Thetermsusersassigntodescribeimagesarealsotheonestheyaremostlikelytotypewhilesearchingfordigitalimages.Likewise,accordingtoWalsh,a
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
32
studyconductedbytheUniversityofAlbertafoundmorethanfortypercentofcollectionsreviewedusedalocallydevelopedclassificationforindexingandsearchingtheircollections,andmanyoftheseschemescouldworkwellforsearcheswithinthecollectionbyuserswhoarefamiliarwiththecultureofthecollection.28
Usage-dataanalysiscanconstituteusefulinformationthatguidesdecisionsforbuildingdigitalcollectionsthatbetterserveuserneeds.Itcanidentifyalibrary’sdigitalcollections’usersandwhattheywant.Theseareimportantconsiderationstokeepinmindiflibraryservicesaretobeallaboutengagingandbuildingrelationshipwiththeusers.29HarvestingtoanationalportalsuchastheDPLAisbeneficialforHunterLibrary’scollections.Atthesametime,theLibrary’sinstitution-specificcollectionsreceivemorereturnvisits,likelybecauseofsustainedinterestfromthelargeuserbaseoftheuniversity’sstudentsandemployees,anassessmentsupportedbysurveyfindings.Conversely,collectionsnotsodirectlytiedtotheinstitutionreceivethemostone-timeitemviewsandreferrals.Itemsthatgetusedareagoodindicationofwhatuserswantand,asthispaperdemonstrates,thefocusofacademicdigitallibrarycollectionsshouldconsidertheneedsofboththeuniversityaudienceandthegeneralpublic.
REFERENCES
1. Alandingpagereferstothehomepageofacollection.
2. TheDPLAprovidesasingleportalforaccessingdigitalcollectionsheldbyculturalheritageinstitutionsacrosstheUnitedStates.“History,”DigitalPublicLibraryofAmerica,accessedMay19,2016,http://dp.la/info/about/history/.
3. PaulBetty,“AssessingHomegrownLibraryCollections:UsingGoogleAnalyticstoTrackUseofScreencastsandFlash-BasedLearningObjects,”JournalofElectronicResourcesLibrarianship21,no.1(2009):75–92,https://doi.org/10.1080/19411260902858631.
4. Ibid.
5. WeiFang,“UsingGoogleAnalyticsforImprovingLibraryWebsiteContentandDesign:ACaseStudy,”LibraryPhilosophyandPractice(e-journal),June2007,1-17,http://digitalcommons.unl.edu/libphilprac/121.
6. MarkBaggettandRabiaGibbs,“HistorypinandPinterestforDigitalCollections:MeasuringtheImpactofImage-BasedSocialToolsonDiscoveryandAccess,”JournalofLibraryAdministration54,no.1(2014):11–22,https://doi.org/10.1080/01930826.2014.893111.
7. MelanieSchlosserandBrianStamper,“LearningtoShare:MeasuringUseofaDigitizedCollectiononFlickrandintheIR,”InformationTechnologyandLibraries31,no.3(September2012):85–93,https://doi.org/10.6017/ital.v31i3.1926.
INFORMATIONTECHNOLOGYANDLIBRARIES|DECEMBER2016 33
8. MarkR.O’English,“ApplyingWebAnalyticstoOnlineFindingAids:PageViews,Pathways,and
LearningaboutUsers,”JournalofWesternArchives2,no.1(2011):1–12,http://digitalcommons.usu.edu/westernarchives/vol2/iss1/1.
9. MarcusLadd,“AccessandUseintheDigitalAge:ACaseStudyofaDigitalPostcardCollection,”NewReviewofAcademicLibrarianship21,no.2(2015):225–31,https://doi.org/10.1080/13614533.2015.1031258.
10.IreneM.H.Herold,“DigitalArchivalImageCollections:WhoAretheUsers?”Behavioral&SocialSciencesLibrarian29,no.4(2010):267–82,https://doi.org/10.1080/01639269.2010.521024.
11.MarkA.MatienzoandAmyRudersdorf,“TheDigitalPublicLibraryofAmericaIngestionEcosystem:LessonsLearnedAfterOneYearofLarge-ScaleCollaborativeMetadataAggregation,”in2014ProceedingsoftheInternationalConferenceonDublinCoreandMetadataApplications(DCMI,2014),1–11,http://arxiv.org/abs/1408.1713.
12. OskanaL.Zavalinaetal.,“ExtendedDate/TimeFormat(EDTF)intheDigitalPublicLibraryofAmerica’sMetadata:ExploratoryAnalysis,”ProceedingsoftheAssociationforInformationScienceandTechnology52,no.1(2015),1–5,http://onlinelibrary.wiley.com/doi/10.1002/pra2.2015.145052010066/abstract.
13. LisaGregoryandStephanieWilliams,“OnBeingaHub:SomeDetailsbehindProvidingMetadatafortheDigitalPublicLibraryofAmerica,”D-LibMagazine20,no.7/8(July/August2014):1–10,https://doi.org/10.1045/july2014-gregory.
14.KateBoyd,HeatherGilbert,andChrisVinson,“TheSouthCarolinaDigitalLibrary(SCDL):WhatIsItandWhereIsItGoing?”SouthCarolinaLibraries2,no.1(2016),http://scholarcommons.sc.edu/scl_journal/vol2/iss1/3.
15.ChrisFreelandandHeatherMoulaison,“DevelopmentoftheMissouriHub:PreparingforLinkedOpenDatabyContributingtotheDigitalPublicLibraryofAmerica,”ProceedingsoftheAssociationforInformationScienceandTechnology52,no.1(2015):1–4,http://onlinelibrary.wiley.com/doi/10.1002/pra2.2015.1450520100105/abstract.
16.Asingleviewofaniteminadigitalcollection.
17.Visitstothesitethatbeganfromanothersitewithanitempagebeingthefirstpageviewed.
18.KeywordsarewordsvisitorsusedtofindtheLibrary’swebsitewhenusingasearchengine.GoogleAnalyticsprovidesalistofthesekeywords.
19.Asessionisdefinedasa“groupofinteractionsthattakeplaceonawebsitewithinagiventimeframe”andcanincludemultiplekindsofinteractionslikepageviews,socialinteractions,andeconomictransactions.InGoogleAnalytics,asessionbydefaultlaststhirtyminutes,though
ANALYZINGDIGITALCOLLECTIONSENTRANCES:WHATGETSUSEDANDWHYITMATTERS|BISWASANDMARCHESONI|https://doi.org/10.6017/ital.v35i4.9446
34
onecanadjustthislengthtolastafewsecondsorseveralhours.“HowaSessionIsDefinedinAnalytics,”Google,AnalyticsHelp,accessedMay20,2016,https://support.google.com/analytics/answer/2731565?hl=en.
20.Locationswerestudiedintermsofmostlycitiesandstates.
21.Thepercentageisbasedonthetotalreferralcountacollectiongets—forexample,a44percentreferralcountforCherokeeTraditionswouldmeanthatthesearchenginesaccountfor44percentofthetotalreferralsthiscollectiongets.
22.Herold,“DigitalArchivalImageCollections,”278.
23.KrystynaK.Matusiak,“TowardsUser-centeredIndexinginDigitalImageCollections,”OCLCSystems&Services:InternationalDigitalLibraryPerspectives22,no.4(2006):283–98,https://doi.org/10.1108/10650750610706998.
24.Ladd,“AccessandUseintheDigitalAge,”230.
25.FangpointsoutthattheimprovementsmadetotheRutgers-NewarkLawLibrarywebsitecouldattractmorereturnvisitorsandthusachieveloyalty.Fang,“UsingGoogleAnalyticsforImprovingLibraryWebsite,”11.
26.NISOFrameworkAdvisoryGroup,AFrameworkofGuidanceforBuildingGoodDigitalCollections,2nded.(Bethesda,MD:NationalInformationStandardsOrganization,2004),https://chnm.gmu.edu/digitalhistory/links/cached/chapter3/link3.2a.NISO.html.
27.Matusiak,“TowardsUser-centeredIndexing,”289.
28.JohnWalsh,“TheUseofLibraryofCongressSubjectHeadingsinDigitalCollections,”LibraryReview60,no.4(2011),https://doi.org/10.1108/00242531111127875.
29.LynnSilipigniConnaway,TheLibraryintheLifeoftheUser:EngagingwithPeopleWhereTheyLiveandLearn,(Dublin:OCLCResearch,2015),http://www.oclc.org/research/publications/2015/oclcresearch-library-in-life-of-user.html.
Reproduced with permission of the copyright owner. Further reproduction prohibited withoutpermission.