SlideShare a Scribd company logo
Semi-­‐Automa+c	
  Visual	
  
Explora+on	
  of	
  Linked	
  Data	
  
Le;eris	
  Tatakis	
  
Le;eris.tatakis@manchester.ac.uk	
  
@Ltatakis	
  
Overview	
  of	
  Problem	
  
hEp://blog.datalicious.com/data-­‐visualisa+on/phone-­‐data-­‐malte-­‐spitz-­‐die-­‐zeit-­‐germany/	
  
mapping	
  
What	
  we	
  have	
   Where	
  we	
  want	
  to	
  be	
  
Overview	
  Example	
  
•  Open	
  Data	
  Manchester	
  Linked	
  Data	
  visualisa+ons.	
  
•  You	
  need	
  to	
  specify:	
  SPARQL	
  Query	
  and	
  Visualisa+on.	
  
•  Requires:	
  +me,	
  effort	
  and	
  technical	
  knowledge.	
  
Heat	
  map	
  of	
  fires	
   Street	
  light	
  issues	
  
hEps://github.com/GMDSP-­‐Linked-­‐Data/PhaseTwo/tree/master/Visualisa+ons	
  
Research	
  Goals	
  
•  Our	
  aim	
  is	
  to	
  support	
  developers	
  who	
  provide	
  
tools	
  for	
  end	
  users.	
  
•  Allow	
  the	
  developers	
  to	
  explore,	
  evaluate	
  and	
  
understand	
  underlying	
  data.	
  
•  “I	
  have	
  all	
  this	
  data	
  but	
  I	
  don’t	
  know	
  what	
  
informa+on	
  is	
  in	
  here!”	
  
•  Allow	
  developers	
  to	
  provide	
  their	
  feedback	
  
during	
  the	
  processing	
  of	
  the	
  data	
  in	
  an	
  semi-­‐
automa+c	
  fashion.	
  
Cycle	
  of	
  Visual	
  Analysis	
  
Cycle	
  of	
  visual	
  analysis,	
  BioVis	
  2015,	
  Tableau	
  
Research	
  Ques+ons	
  
•  What	
  are	
  the	
  key	
  characteris+cs	
  that	
  affect	
  the	
  
choice	
  of	
  data	
  visualisa+ons?	
  
•  Which	
  aspects	
  of	
  the	
  user's	
  task	
  must	
  we	
  
understand	
  to	
  provide	
  an	
  appropriate	
  selec+on	
  
of	
  data	
  visualisa+ons?	
  
•  How	
  do	
  we	
  map	
  the	
  characteris+cs	
  of	
  the	
  data	
  
sources	
  to	
  the	
  characteris+cs	
  of	
  exis+ng	
  
visualisa+ons.	
  	
  
ODM	
  Community	
  
•  Do	
  you	
  have	
  Linked	
  Data	
  and	
  have	
  no	
  idea	
  how	
  
to	
  understand/explore	
  it?	
  	
  
•  I	
  would	
  like	
  to	
  understand	
  the	
  process	
  and	
  the	
  
way	
  you	
  work	
  with	
  the	
  data.	
  
•  How	
  can	
  we	
  improve	
  it?	
  
•  Are	
  visualisa+ons	
  always	
  appropriate?	
  
•  Collabora+ng	
  with	
  the	
  community.	
  	
  
•  Data 	
   	
   	
  	
  
•  Comments	
  
•  Insights	
  	
  
•  Ideas 	
   	
   	
  	
  
•  Interested	
  /	
  Curious	
  	
  
•  Solved	
  it?	
  
Thank	
  you!	
  
@LTatakis	
  
le;eris.tatakis@manchester.ac.uk	
  

More Related Content

Open Projector Open Data Manchester Visualisation Meeting

  • 1. Semi-­‐Automa+c  Visual   Explora+on  of  Linked  Data   Le;eris  Tatakis   Le;eris.tatakis@manchester.ac.uk   @Ltatakis  
  • 2. Overview  of  Problem   hEp://blog.datalicious.com/data-­‐visualisa+on/phone-­‐data-­‐malte-­‐spitz-­‐die-­‐zeit-­‐germany/   mapping   What  we  have   Where  we  want  to  be  
  • 3. Overview  Example   •  Open  Data  Manchester  Linked  Data  visualisa+ons.   •  You  need  to  specify:  SPARQL  Query  and  Visualisa+on.   •  Requires:  +me,  effort  and  technical  knowledge.   Heat  map  of  fires   Street  light  issues   hEps://github.com/GMDSP-­‐Linked-­‐Data/PhaseTwo/tree/master/Visualisa+ons  
  • 4. Research  Goals   •  Our  aim  is  to  support  developers  who  provide   tools  for  end  users.   •  Allow  the  developers  to  explore,  evaluate  and   understand  underlying  data.   •  “I  have  all  this  data  but  I  don’t  know  what   informa+on  is  in  here!”   •  Allow  developers  to  provide  their  feedback   during  the  processing  of  the  data  in  an  semi-­‐ automa+c  fashion.  
  • 5. Cycle  of  Visual  Analysis   Cycle  of  visual  analysis,  BioVis  2015,  Tableau  
  • 6. Research  Ques+ons   •  What  are  the  key  characteris+cs  that  affect  the   choice  of  data  visualisa+ons?   •  Which  aspects  of  the  user's  task  must  we   understand  to  provide  an  appropriate  selec+on   of  data  visualisa+ons?   •  How  do  we  map  the  characteris+cs  of  the  data   sources  to  the  characteris+cs  of  exis+ng   visualisa+ons.    
  • 7. ODM  Community   •  Do  you  have  Linked  Data  and  have  no  idea  how   to  understand/explore  it?     •  I  would  like  to  understand  the  process  and  the   way  you  work  with  the  data.   •  How  can  we  improve  it?   •  Are  visualisa+ons  always  appropriate?   •  Collabora+ng  with  the  community.     •  Data         •  Comments   •  Insights     •  Ideas         •  Interested  /  Curious     •  Solved  it?  
  • 8. Thank  you!   @LTatakis   le;eris.tatakis@manchester.ac.uk