adb -使數(shù)據(jù)易于獲取：利用統(tǒng)計(jì)數(shù)據(jù)和元數(shù)據(jù)交換 Making Data Easily Accessible - Leveraging Statistical Data and Metadata eXchange

NO.336

FEBRUARY2025

ADBBRIEFS

KEYPOINTS

?StatisticalDataandMetadataeXchange(SDMX)isan

ISOstandardaimedat

definingandstandardizing

theexchangeofstatistical

dataandmetadata,aswell

asenhancingtheefficient

sharingofinformationamongvariousorganizations.

?AdoptionofSDMXimprovesdataqualityandefficiency.

Theimplementationof

SDMXreducesmanualdataprocessinganderrorswhilepromotingautomationin

datavalidation,processing,andreporting.

?SDMXisrapidlygaining

momentum.Itisfast

becomingadrivingforceinstreamliningdataactivitiesandpromotingstatistical

interoperability,withover

100nationalstatisticsofficesandcentralbanksalready

usingthestandard.

?SDMXhasbroad

applicability.SDMX

principlesarealsoapplicabletoawiderangeofnonofficialstatisticalindicators

publishedbygovernmentsandinternational

organizations.

?Capacitybuildingis

essentialtosuccessful

SDMXimplementation.Itisimportanttoprovide

accessibleandcost-efficienttrainingtoequipdata

producersanduserswiththeskillstoapplythestandard

effectively.

ISBN978-92-9277-176-8(print)ISBN978-92-9277-177-5(PDF)

ISSN2071-7202(print)ISSN2218-2675(PDF)

PublicationStockNo.BRF250028-2

DOI:

/10.22617/BRF250028-2

MakingDataEasilyAccessible:LeveragingStatisticalData

andMetadataeXchange

BrianBuffettConsultant

AsianDevelopmentBank(ADB)

StefanSchipperSeniorStatisticianADB

JeffreyNapolesConsultant

ADB

PamelaLapitanStatisticsOfficerADB

THEORIGINSOFFUNDAMENTALSTATISTICALPRINCIPLES

Statisticsplayacriticalroleinshapingdevelopmentinitiatives—aviewsharedbymanyplanners,policymakers,researchers,dataanalysts,andstatisticiansinnationaland

internationalinstitutions,aswellasamongacademics.

Theneedforguidingprinciplesforofficialstatisticsbecameevidentinthelate1980s

asCentralEuropetransitionedfromcentrallyplannedeconomiestomarket-driven

democracies.Inresponse,theConferenceofEuropeanStatisticiansdevelopedand

endorsedtheFundamentalPrinciplesofOfficialStatisticsin1991.Itsoonbecame

clear,however,thattheseprincipleshadbroaderrelevancebeyondEurope.Afteran

internationalconsultationprocess,theUnitedNationsStatisticalCommission(UNSC)adoptedtheseprinciplesonaglobalscale.Witharevisedpreamble,theUnitedNationsFundamentalPrinciplesofOfficialStatisticswereformallyadoptedinApril1994and

laterendorsedbytheUnitedNationsGeneralAssemblyin2014,markingasignificantmilestoneintheevolutionofinternationalstatistics.

Note:TheworkhasbenefitedfromcommentsreceivedfromElaineS.Tan,DirectoroftheDataDivisionoftheEconomicResearchandDevelopmentImpactDepartment,AsianDevelopmentBankandPaul

Dent(editor).

2

ADBBRIEFSNO.336

Tohelpstatisticalorganizationsgaugehowwelltheyhave

operationalizedthefundamentalprinciples,aself-assessmentmaturitymodelwasproducedfortheUNSCin2020.

Selecttargetsforhighstatisticalmaturityunderthismodelareasfollows:

(i)Usershaveunrestrictedaccesstoexploreandutilizedata,includingthroughonlineplatforms.

(ii)Datadownloadsareofferedinopen,machine-readable

formatssuchasExtensibleMarkupLanguage(XML),

JavaScriptObjectNotation(JSON),andcomma-separatedvalues(CSV).Onlineapplicationprogramminginterfaces

enabledirectinteractionwithstatisticaloutputs.

(iii)Releasedatesandtimesarepubliclyannouncedinadvance.Anychangestoreleaseschedulesarepublished,alongwithdocumentedexplanations.

(iv)Statisticalmethodologiesadheretointernationally

recommendedstandardsandpracticewhereavailable.

(v)Qualityreportsaccompanydatareleases,anduserguidesareprovidedtohelpinterpretthedataeffectively.

(vi)Usersareinformedofanymethodologicalchangesor

correctionstoerrors.Transparentprocessesandpubliclyavailablepoliciesguidethereportingofsuchupdates.

(vii)Ensuringcoherenceandconsistencyacrossstatisticalsystemsisastrategicpriority.Governancemechanismsoverseeandpromotealignmentacrossdatasources.

(viii)Statisticalconceptsandresourcesareharmonizedacross

producerswheneverpossible.Centralizedresourcessupportthealignmentofoutputamongproducers.

(ix)Nationalproducerbodiescollaboratetoensurecoherenceandadoptaunifiedapproachtostatisticalcoordination.

(x)Internationallyrecognizedconcepts,classifications,and

methodsareusedfordeveloping,producing,anddisseminatingofficialstatistics.Newversionsareadoptedasappropriate.

(xi)Theuseofinternationalstandardsisactivelyencouraged

withinthenationalstatisticalsystem,witheffortstoengageinthedevelopmentofthesestandards.

Asofthistime,achievingthesematuritytargetsremainsan

aspirationformanystatisticalorganizations—anaspirationthatisunlikelytoberealizedwithoutadeeperanalysisofhowthe

organizationsthemselvesareworking.

ESSENTIALBENEFITSOFSTATISTICALDATAANDMETADATAEXCHANGE

StatisticalDataandMetadataeXchange(SDMX)isan

InternationalOrganizationforStandardization(ISO)standardthathasemergedasameanstostreamlinedataactivitiesandpromotestatisticalinteroperabilityacrosstheworld.

SDMXwasoriginallydevelopedtofocusonstatistical

macro-aggregatesbuthassinceexpandedtosupportmicrodataandvariousunstructureddataformats.Althoughitwasdesignedwithofficialstatisticsasitsprimaryfocus,SDMXoffersthe

flexibilitytobeappliedacrossawiderangeofdomains.

Developedinconsultationwithnationalstatisticsoffices(NSOs),theSDMXstandardprovidesacomprehensiveframeworkfor

structuring,collecting,producing,exchanging,andmanagingstatisticaldataandmetadata.Bydoingso,SDMXenables

seamlessintegrationandsharingofdataacrossdifferentstatisticalsystemsanddomains.Itisconservativelyestimatedthatover

100NSOsandcentralbankshaveadoptedtheSDMXstandardandassociatedopen-sourcetoolstoimprovetheirstatistical

productionprocessesandmethods.

TherearemanybenefitsofSDMXthatcanassistorganizationsintheireffortstoachievetargetsundertheUNSC’smaturitymodelfortheFundamentalPrinciplesofOfficialStatistics.Forexample,adoptionoftheSDMXstandardfacilitatestheproductionof

better-qualityandmoretimelydata.

SDMXensuresdataaccuracythroughprevalidation,automatingthereportingprocess,standardizingdatastructures,andenablingefficientdatacollectionmethods.Thestandardallowsforthe

prevalidationofdatamessagesagainstpredefineddefinitions.

Thisensuresthatthedatamessagesarealigned,reducingerrorsduringdataexchangeandminimizingtheneedformanual

intervention.Asaresult,dataproducerscanchecktheaccuracyandadherenceoftheirdatasetsbeforesendingthem.Byavoiding

errorsandensuringcomplianceatthesource,theSDMX

prevalidationfeatureminimizestheneedformultipleroundsofcommunicationbetweendataproducersandconsumers

torectifydataissues.Itstreamlinesthereportingprocessandreducestheback-and-forthexchangesthatoftenoccurduringdatavalidation.

SDMXpromotestimelinessbyreducingtheneedformanual

dataconversionandfacilitatingautomatedchecksforswiftdatavalidation.Automatedvalidationprocessescanquicklyidentify

errorsorinconsistencies,reducingthetimeneededformanual

verificationandvalidation.Forexample,whenreportingonthe

percentageofwomeninParliament,anautomatedcheckensuresthatthevalueforthe“sex”fieldiscorrectlysetto“female”

andpreventsanyerroneousentryof“male”or“other”.Such

automationresultsinvalidateddatabeingavailablemorequicklyandletsusersaccessthishigh-qualitydatainashortertime.

SDMXalsoenablesautomatedprocessingofdata,reducingthelikelihoodofhumanerrorsassociatedwithmanualhandling.

Automatedworkflowsensureconsistencyandaccuracyindataprocessing,whichultimatelyimproveoveralldataquality.

Furthermore,adoptionoftheSDMXstandardtriggersincreasedautomationwithinnationalstatisticalsystems.ThisautomationisachievedbytheSDMXapplicationprogramminginterface

(API),whichprovidesprogrammaticaccesstostatisticaldataandmetadata.Forexample,APIdataqueriesallowtheretrievalofstatisticalinformation,rangingfromentiredatasetsto

individualobservations.Similarly,APIstructurequeriesprovideaccesstostructuralmetadataatvariouslevelsofgranularity,

e.g.,fromthefullsetofmetadataavailableinthesourcetoasinglecode.

3

MakingDataEasilyAccessible

WHYSTANDARDIZATIONISIMPORTANTTOSTATISTICALMATURITY

ThepowerbehindSDMXliesinitsinformationmodel,which

allowsforthedevelopmentofprocessesandfunctionsaroundthedatamodel,ratherthanbeingconstrainedbyspecific

datasyntaxesorformats.TheSDMXInformationModel

establishesacommonlanguageandstructureforexpressing

statisticalconceptstoensureconsistencyanduniformityintherepresentationofdata.

SDMXsupportscommonterminologyfordescribingstatistical

data,therebyharmonizingconceptsandcodelists.Thisstrong

terminologicalfoundationstandardizesdatadescriptionacross

variousstatisticaldomains.Harmonizingstatisticaldatacontent

andstructureoffersnumerousadvantages,includingashared

languageamongimplementersandusers.Thisisachievedusing

uniformcodes,withnamesanddescriptionsthatcanbeexpressedindifferentlanguages.

Theterminologicalapproachsavestimeandresourcesthrough

reducedmappinganddataprocessing,wideravailabilityoftoolsbasedonacommonlyagreedformat,andtheexistenceofSDMXregistriesthatfacilitatereuse.

SDMXalsoenhancesinterpretabilitybyusingastandardizedterminologytoharmonizemetadata.Thiscontributestothedevelopmentofaglobalstatisticallanguageandimproves

coherencethroughcross-domainconcepts,sharedcontent,harmonizedstatisticalguidelines,andextensivereuseacrossdomainsandagencies.

TheseobjectivesareamongtheUNSC’skeytargetsforhighmaturityintheFundamentalPrinciplesofOfficialStatistics.

HOWTHESTANDARDPROMOTESINTERAGENCYCOLLABORATION

ThecoreresponsibilitiesofNSOsincludecollecting,processing,anddisseminatingstatisticaldata.Eachorganizationthat

requestsdatafrom,orprovidesdatato,anNSOmayhaveitsownuniquedataformatrequirements.Intheseinstances,theNSO

mayfinditselfhavingtoadaptandtransformthesamedatasetintomultipleformats,(e.g.,spreadsheet,textfile,digitalfile),

significantlyincreasingitsworkloadandresourcerequirements.AdoptionoftheSDMXstandardproveshighlybeneficialin

resolvingthesescenarios.

SDMX’sstandardizedapproachenablesinteroperablefunctionalitywithinandbetweensystemsdedicatedtoexchanging,reporting,

anddisseminatingstatisticalinformation.Theresultisimproveddataandmetadataconsistencyandcomparabilityacrossdifferentstatisticalapplicationsandorganizations.

BecauseSDMXpromotestheuseofstandardizeddatastructuresandexchangeformats,organizationscanaligntheirreporting

systemswiththesestructuresandformats,reducingtheneedformaintainingmultiple,diversereportingsystems.Standardizationacrossorganizationsanddatadomainsfurtherminimizesthe

complexityassociatedwithmanagingvariousreportingformatsandexchangeagreements.Thisstreamliningcontributestoa

significantreductionintheoveralldatareportingburden.

SDMXisaversatileandaccessibleframework,providingflexibilityforcustomizationandallowingcost-efficienciesviaaccesstofreelyavailabledevelopmenttoolsandcodes.MostSDMXtechnology

toolsarefreeandopen-source,whichnotonlyreducesthecost

ofacquiringorproducingtoolsbutalsoallowsorganizationsto

benefitfromawiderangeofsolutionscreatedandmaintainedbytheSDMXcommunity.Additionally,theopen-sourcecommunitysharesexpertiseandbestpractices,providingvaluableresources

atnocost.Thesefactorsgoalongwaytowardremovingbarrierstoimplementationanddataaccessibility.

SDMXisalsodesignedtobeflexibleandinteroperablewith

moderntechnologies.ThisadaptabilityallowsNSOsandother

data-producingagenciestointegrateSDMXwithcontemporaryITinfrastructures,databases,anddata-processingtools,promotingamoremodernandefficientdataecosystem.

Organizationscanthereforecustomizetheirimplementation

ofSDMXtoalignwiththeirspecificneeds,takingintoaccount

theirexistinginfrastructure,dataworkflows,andoperational

priorities.Thisflexibilityallowsthemtoadaptthestandardtosuittheiruniquedataproductionprocesses,technicalcapabilities,andorganizationalgoals.WhileNSOsareamongtheprimaryusers

ofSDMX,otherorganizations,suchasinternationalagencies,researchinstitutions,andprivatesectorsorganizationscan

alsoleverageitsadaptabilitytosupportdiversedata-sharingneed.

AdoptionoftheSDMXstandardalsocontributestothe

modernizationofstatisticaldatamanagementpracticesworldwide.

Becausethestandardisgloballyendorsedandadopted,it

encouragescollaborationamonginternationalorganizationsandnationaleconomiesandfosterstheexchangeofmodernpractices,methodologies,andtechnologies,creatingasharedandregularlyupdatedvisionforstatisticaldatamanagement.

THEROLEOFTHESTANDARDIN

DATAPRODUCTIONANDSHARING

TheSDMXstandardbringsmanybenefitstostatisticalbusiness

processes,includingdatacollection,production,reporting,and

dissemination.Importantly,useofthestandardmakesiteasiertoaggregate,compare,andanalyzeinformationfromdiversesources.This,inturn,makesitmoreefficientfororganizationstoshare

andusereadilyunderstoodstatisticalinformationatnationalandinternationallevels.

4

ADBBRIEFSNO.336

Harmonizingdataandmetadatafromdiversesourcesinvolvesaligningconcepts,classifications,andstructurestoensure

consistency.Lackofharmonizationcanresultindifficultiesin

comparingandaggregatingdataacrossdifferentstatisticaldomainsandmaylimittheusefulnessofstatisticalinformation.Inaddition,withoutharmonization,thesharingofstatisticaldatabecomes

morechallengingandlessefficient.

SDMXsupportsthecomplexandhighlyspecializedtasks

ofvalidatingandaggregatingdatawiththeintroduction

ofmetadata-drivenprocessesandtools.Inthiscontext,

ametadata-drivenprocessisasystematicapproachto

modelling,structuring,andorganizingtheinformationabout

data.Thebenefitofsuchaprocessincreasesevenfurtherif

validation,transformation,andaggregationprocessesaredrivenbymetadatathatidentifyanddescribedata.

SDMXmakesclearstatisticalinformationavailabletousersby

definingstandardizedstructuresforthemetadatausedtodescribethecontentandqualityofstatisticaldata.Suchdefinitions

mayincludethedata’ssources,methodologies,andquality

indicators.Thismetadatahelpsusersunderstandthecontext

andcharacteristicsofthedatabeingdisseminated.Additionally,theSDMXstandardenablesthepublicationofdatainaformatthatiseasilyaccessibleandunderstandable.

DATAGOVERNANCE

UNDERTHESTANDARD

TheSDMXstandardwasdevelopedandissponsoredbyeightmajorinternationalorganizations(asshowninthefigurebelow)andoperatesunderawell-establishedgovernancemodel.

TechnicalWorkingGroup

(TWG)

StatisticalWorkingGroup

(SWG)

GovernanceStructurefortheStatisticalDataandMetadataeXchangeInitiative

SponsorOrganizations

BIS,ECB,EUROSTAT,IMF,OECD,

UN,WORLDBANK,ILO

SDMXSponsorsCommittee

SDMXSecretariat

InformationModel

SDMXGlobalRegistry

SDMXO代cialWebsite

BIS=BankforInternationalSettlements,ECB=EuropeanCentralBank,Eurostat=StatisticalOfficeoftheEuropeanUnion,ILO=InternationalLabourOrganization,IMF=InternationalMonetaryFund,OECD=OrganisationforEconomicCo-operationandDevelopment,SDMX=StatisticalDataand

MetadataeXchange,UN=UnitedNations.

Source:AsianDevelopmentBankvisualizationbasedon

/?page_id=2561

.

5

MakingDataEasilyAccessible

KeycomponentsofthisgovernancestructureincludetheSponsors

Committee,whichoverseesthestrategicdecision-making,

andtheSecretariat,responsiblefortheinitiative’soperational

management.Thishierarchyensuresefficientcoordination

anddecision-makingamongthesponsoringorganizations.In

addition,theTechnicalWorkingGroupandStatisticalWorking

GroupplayaproactiveroleincontinuouslyimprovingtheSDMXstandardtomeettheneedsofusers,ensuringitremainsboth

relevantandadaptable.

Anothercriticalaspectofdatagovernanceunderthestandard

istheSDMXInformationModel.Itdescribesthekeyconcepts

aroundstatisticaldata,metadata,anddataexchangeprocesses.

Themodelcanbeusedtodescribeanymultidimensionaldataset,regardlessofstatisticaldomain,andistheareaofprimaryfocus

forstatisticians.ThemodelservesasthecentralandfundamentalcomponentoftheSDMXframework,definingthestructureof

statisticaldataandmetadata.

Morespecifically,themodelidentifiesobjectswithinthestatisticaldomainanddefinestheirrelationships.Thisincludesessential

elementssuchascoreconcepts,theirroles,andthecodeliststhatenableaclearunderstandingoftheinterconnectionsbetween

variousstatisticalagencies.

TheSDMXregistrysupportstheseeffortsbyhelpingstatisticalorganizationsexternalizeandgaincontroloftheirstatistical

metadata.TheSDMXGlobalRegistryservesasacentralizedrepository,housingglobalstructures,concepts,andcodelistsdefinedbyexpertsacrossvariousstatisticaldomains.

Organizationscanleveragetheregistryasavaluableresourceforguidanceonstructuring,processing,validating,andinterpretingstatisticaldata.

TheSDMXContent-OrientedGuidelinesprovideproceduresandbestpracticesforcreatingandmanagingthecontentofstatisticaldataandmetadatawithintheSDMXframework.

Aprimaryfocusoftheguidelinesisharmonizingspecificconceptsandterminologiesthatarecommontoalarge

numberofstatisticaldomains.Bydoingso,theguidelines

createasharedlanguage,contributingtoamoreefficient

exchangeofcomparabledataandmetadata.Thisisparticularlyvaluablewheredatasetsneedtobecomparedorcombined

acrossdifferentstatisticaldomains:commonterminology

andconceptsenhancetheconsistencyandreliabilityofsuch

comparisons.Theguidelinesofferrecommendedpractices

thatcanbeimplementedconsistently,ensuringinteroperabilityandstandardizationregardlessofthespecificfocusorsubjectmatterofthestatisticaldata.Theyfurthersupportthecreationofinternationalgoodpracticesandsharedstandards,suchas

domain-specificdatamodelsandcross-domaincodelists.

Theyarealsodevelopedusingimplementationexperiences

withintheSDMXcommunity,leveraginglessonslearnedfrompracticalimplementationstoensurethatallguidelinesare

groundedinreal-worldscenariosandarereflectiveofbestpracticesglobally.

SUPPORTINGADOPTIONVIAFREELYAVAILABLELEARNINGRESOURCES

OneofthekeychallengesofadoptingSDMXwithinnational

statisticalsystemsisalackofknowledgeonthestandardandtherelevanttechnologiesandtoolsassociatedwithit.

Sinceknowledgegapsexistacrossallmanageriallevels,itiscrucialtooffertrainingprogramsandworkshopsthatcancatertoawidevarietyofstaffcohorts,frombeginnerstomoreadvancedusers.

Suchprogramsensurethatparticipantscanbeequippedwith

anunderstandingofthefoundationalconceptsofSDMXand/orthemoretechnicalaspectsofthestandard.

Tosupportthiscapacitybuildingwithinstatisticalorganizations,severalonlinetrainingprogramshavebeendeveloped,allowingparticipantstolearnattheirownpaceandbasedontheirlevelofexpertise.Moreover,manyofthee-learningcoursesare

offeredfreeofcharge,reducingfinancialbarriersforlearnersandorganizationsalike.

TheAsianDevelopmentBankhascollaboratedwithdevelopmentpartnerstoreleaseandconducttwovitalSDMXe-learning

trainingprograms.Participantscanchoosetoenrollinthe

SDMXFoundationCourse1and/ortheSDMXToolsCourse2

dependingontheiraptitudesandlearningrequirements.More

than700peoplehavealreadysuccessfullycompletedandbeen

certifiedintheSDMXFoundationand/orSDMXToolsCourse,

withmorethan95%ofthegraduatesratingthesecoursesaseitherexcellentorgood.

Inaddition,avarietyofe-learningcoursesandguidelinesonSDMXareavailableontheofficialSDMXwebsite.3These

resourcesprovidecomprehensivesupportforlearnersandpractitioners,offeringdetailedguidanceandbestpracticesonimplementingtheSDMXstandard.

1AccesstheSDMXFoundatione-learningcoursevia

/course/view.php?id=486

.Accesstheuserguidevia

/mod/book/

view.php?id=10142&chapterid=2138

.

2AccesstheSDMXToolse-learningcoursevia

/course/view.php?id=520

.

3Accessthesee-learningcoursesontheLearningResourcespage(

/?page_id=2555

)andexploreguidelinesandbestpracticesmaterialsunderGuidelinessection(

/?page_id=4345

)oftheofficialSDMXwebsite.

6

ADBBRIEFSNO.336

CONCLUSION

Severalmaturitytargetsidentifiedatthebeginningofthis

brieffocusondataaccess,interpretation,anduse.TheSDMX

standard—alongwiththeSDMXContent-OrientedGuidelinesandtheSDMXRegistry—offersdefinitivepathwaysforachievingthesematuritytargets.

ManyoftheUNSC’smaturitytargetsfortheFundamental

PrinciplesofOfficialStatisticsalsopertaintodatagovernanceandimprovingcoherence,consistency,anduseofinternationalstandardsandgoodpractices.ThecombinationoftheSDMX

standard,SDMXContent-OrientedGuidelines,SDMXRegistry,andSDMXAPIscansignificantlyimproveoperationalefficienciestowardthesetargets.

REFERENCES

AsianDevelopmentBank.2024.

EnhancingDataManagement

ThroughtheStatisticalDataandMetadataeXc