Data CItation Principles in Practice

933 views

Published on

Talk at IDCC 2014 http://www.dcc.ac.uk/events/idcc14 Data Citation Principles Workshop, Feb 27, 2014

Published in: Entertainment & Humor
2 Comments
1 Like
Statistics
Notes
No Downloads
Views
Total views
933
On SlideShare
0
From Embeds
0
Number of Embeds
30
Actions
Shares
0
Downloads
8
Comments
2
Likes
1
Embeds 0
No embeds

No notes for slide

Data CItation Principles in Practice

  1. 1. Pu#ng  Principles  in  Prac/ce:   Can  a  Publisher  Implement   Data  Cita/on?     Anita  de  Waard   VP  Research  Data  Collabora/ons   a.dewaard@elsevier.com         hCp://researchdata.elsevier.com/      
  2. 2. Can  a  publisher  bring  the  Data  Cita/on   Principles  into  prac/ce?   1.  Importance:  Data  should  be  considered  legi/mate,  citable  products  of  research.  Data  cita/ons  should   be  accorded  the  same  importance  in  the  scholarly  record  as  cita/ons  of  other  research  objects,  such  as   publica/ons.     2.  Credit  and  aCribu/on:  Data  cita/ons  should  facilitate  giving  scholarly  credit  and  norma/ve  and  legal   aCribu/on  to  all  contributors  to  the  data,  recognizing  that  a  single  style  or  mechanism  of  aCribu/on   may  not  be  applicable  to  all  data.   3.  Evidence:  Where  a  specific  claim  rests  upon  data,  the  corresponding  data  cita/on  should  be  provided.   4.  Unique  Iden/fica/on:  A  data  cita/on  should  include  a  persistent  method  for  iden/fica/on  that  is   machine  ac/onable,  globally  unique,  and  widely  used  by  a  community.   5.  Access:  Data  cita/ons  should  facilitate  access  to  the  data  themselves  and  to  such  associated  metadata,   documenta/on,  and  other  materials,  as  are  necessary  for  both  humans  and  machines  to  make  informed   use  of  the  referenced  data.   6.  Persistence:  Metadata  describing  the  data,  and  unique  iden/fiers  should  persist,  even  beyond  the   lifespan  of  the  data  they  describe.   7.  Versioning  and  granularity:  Data  cita/ons  should  facilitate  iden/fica/on  and  access  to  different  versions   and/or  subsets  of  data.  Cita/ons  should  include  sufficient  detail  to  verifiably  link  the  ci/ng  work  to  the   por/on  and  version  of  data  cited.   8.  Interoperability  and  flexibility:  Data  cita/on  methods  should  be  sufficiently  flexible  to  accommodate  the   variant  prac/ces  among  communi/es  but  should  not  differ  so  much  that  they  compromise   interoperability  of  data  cita/on  prac/ces  across  communi/es.    
  3. 3. 1.  Importance:  Data  should  be  considered  legi/mate,  citable  products  of   research.  Data  cita/ons  should  be  accorded  the  same  importance  in  the   scholarly  record  as  cita/ons  of  other  research  objects,  such  as  publica/ons.   hCp://www.sciencedirect.com/science/ar/cle/pii/S1386142513009098  
  4. 4. 2.  Credit  and  aCribu/on:  Data  cita/ons  should  facilitate  giving  scholarly  credit  and   norma/ve  and  legal  aCribu/on  to  all  contributors  to  the  data,  recognizing  that  a   single  style  or  mechanism  of  aCribu/on  may  not  be  applicable  to  all  data.   http://www.sciencedirect.com/science/article/pii/S0370157304002753
  5. 5. 3.  Evidence:  Where  a  specific  claim  rests  upon  data,  the   corresponding  data  cita/on  should  be  provided.   Presen'ng  Supplementary  Material  at  the  relevant  loca'on     •  Supplementary  material  inserted  at   the  place  of  reference/cita/on   •  Put  material  into  the  right  context   •  Make  it  easier  for  readers  to  find   •  Ini/ally  in  closed  text-­‐box,  ac/on  to   open  
  6. 6. Small  side  note:   Taking  evidence  a  step  further:   Cortex  Registered  Report: •  Two-step submission process: •  Method and proposed analysis are submitted for pre-registration •  Paper is conditionally accepted •  Research is executed •  Full paper submitted, accepted provided that protocol is followed •  All experimental data made available Open Access Featured in The Guardian “Confronting the 'sloppiness' that pervades science”, http://bit.ly/1aUAy7f
  7. 7. 4.  Unique  Iden/fica/on:  A  data  cita/on  should  include  a   persistent  method  for  iden/fica/on  that  is  machine  ac/onable,   globally  unique,  and  widely  used  by  a  community.     Interlinking  Ar/cles  and  Data  through  accession  numbers   Enabling one-click access to relevant primary data   •  Author-tagged •  Captured in article XML •  Linked to data repository from the online hCp://public.lanl.gov/herbertv/papers/ article on ScienceDirect Papers/2014/IDCC2014_vandesompel.pdf   See  hCp://www.elsevier.com/databaselinking  
  8. 8. 5.  Access:  Data  cita/ons  should  facilitate  access  to  the  data   themselves  and  to  such  associated  metadata,  documenta/on,  and   other  materials,  as  are  necessary  for  both  humans  and  machines  to   make  informed  use  of  the  referenced  data.     Integrating (meta)data into the article page view   •  Supplementary data at PANGAEA •  Bidirectional links between PANGAEA <> ScienceDirect •  Data visualized next to the article See  hCp://www.elsevier.com/databaselinking  
  9. 9. 6.  Persistence:  Metadata  describing  the  data,  and  unique  iden/fiers   should  persist,  even  beyond  the  lifespan  of  the  data  they  describe.  
  10. 10. 7.  Versioning  and  granularity:  Data  cita/ons  should  facilitate  iden/fica/on   and  access  to  different  versions  and/or  subsets  of  data.  Cita/ons  should   include  sufficient  detail  to  verifiably  link  the  ci/ng  work  to  the  por/on  and   version  of  data  cited.   •  Discussion  with  Dave  DeRoure:     –  How  do  you  reference  a  Research  Object?     –  Is  that  a  good  way  to  describe  an  experiment?   –  (Should  we  start  a  Force11  WG  on  it?)   •  Discussion  with  David  Rosenthal:     –  Are  DOIs  really  the  best  iden/fiers  for  datasets?     –  Perhaps  URI’s  (that  can  have  a  hierarchical  structure,  cf.   DNS)  are  a  beCer  iden/fier  mechanism?   •  Requirement  from  Herbert  van  den  Sompel:   make  it  machine-­‐ac/onable!   •  Ques/on  to  you  all:     –  What  is  a  dataset?  (Cf.  David  Minor:  what  is  an  object?)  
  11. 11. 8.  Interoperability  and  flexibility:  Data  cita/on  methods  should  be  sufficiently   flexible  to  accommodate  the  variant  prac/ces  among  communi/es  but  should   not  differ  so  much  that  they  compromise  interoperability  of  data  cita/on   prac/ces  across  communi/es.     hCp://www.elsevier.com/about/content-­‐innova/on/database-­‐linking  
  12. 12. Can  a  publisher  bring  the  Data  Cita/on   Principles  into  prac/ce?   1.  Importance:  Data  should  be  considered  legi/mate,  citable  products  of  research.  Data  cita/ons  should   be  accorded  the  same  importance  in  the  scholarly  record  as  cita/ons  of  other  research  objects,  such  as   publica/ons.     2.  Credit  and  aCribu/on:  Data  cita/ons  should  facilitate  giving  scholarly  credit  and  norma/ve  and  legal   aCribu/on  to  all  contributors  to  the  data,  recognizing  that  a  single  style  or  mechanism  of  aCribu/on   may  not  be  applicable  to  all  data.   3.  Evidence:  Where  a  specific  claim  rests  upon  data,  the  corresponding  data  cita/on  should  be  provided.   4.  Unique  Iden/fica/on:  A  data  cita/on  should  include  a  persistent  method  for  iden/fica/on  that  is   machine  ac/onable,  globally  unique,  and  widely  used  by  a  community.   5.  Access:  Data  cita/ons  should  facilitate  access  to  the  data  themselves  and  to  such  associated  metadata,   documenta/on,  and  other  materials,  as  are  necessary  for  both  humans  and  machines  to  make  informed   use  of  the  referenced  data.   6.  Persistence:  Metadata  describing  the  data,  and  unique  iden/fiers  should  persist,  even  beyond  the   lifespan  of  the  data  they  describe.   7.  Versioning  and  granularity:  Data  cita/ons  should  facilitate  iden/fica/on  and  access  to  different  versions   and/or  subsets  of  data.  Cita/ons  should  include  sufficient  detail  to  verifiably  link  the  ci/ng  work  to  the   por/on  and  version  of  data  cited.   8.  Interoperability  and  flexibility:  Data  cita/on  methods  should  be  sufficiently  flexible  to  accommodate  the   variant  prac/ces  among  communi/es  but  should  not  differ  so  much  that  they  compromise   interoperability  of  data  cita/on  prac/ces  across  communi/es.    

×