Transcript
Page 1: PATHS at Digital Humanities Congress 2012

Navigating Cultural Heritage Collections using Pathways

N. Aletras, P.D. Clough, S. Fernando, N.Ford, P. Goodale, M.M. Hall, M. Stevenson

University of Sheffield

Information School / Department of Computer Science

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Page 2: PATHS at Digital Humanities Congress 2012

Opening Up Digital Cultural Heritage

Digital Humanities Congress 2012, Sheffield, 6th – 8th Septemberhttp://www.flickr.com/photos/usnationalarchives/4069633668/

Carl Collinshttp://www.flickr.com/photos/carlcollins/199792939/

http://www.flickr.com/photos/brokenthoughts/122096903/

Page 3: PATHS at Digital Humanities Congress 2012

Opening Up Digital Cultural Heritage

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

http://www.flickr.com/photos/28481088@N00/3731283554/

Page 4: PATHS at Digital Humanities Congress 2012

Cutting Paths through DCH

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

http://www.flickr.com/photos/28481088@N00/3731283554/

Leeds Town Hall

Leeds City Art Gallery – Entrance

Leeds City Art Gallery - Interior

On The Headrow, sandwiched between the Central Library and Henry Moore Institute. The grand interior houses a notable collection of fine art and the cafe in the recently renovated Victorian Tiled Hall is well worth a visit. The unremarkable exterior architecture is overshadowed by a large broze sculpture by Henry Moore.

Page 5: PATHS at Digital Humanities Congress 2012

Cutting Paths through DCH

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

http://www.flickr.com/photos/28481088@N00/3731283554/

Follow Paths

Explore the Collection

Share their own Paths

Page 6: PATHS at Digital Humanities Congress 2012

Our Collections

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Collection Language Number & Type of ItemsCulture Grid English 547,741 ImagesHispana Spanish 1,129,640 Texts

105,493 ImagesCervantes Virtual Spanish 19278 Texts

Page 7: PATHS at Digital Humanities Congress 2012

Filtering the Data

• Some items have very limited meta-data• Tried filtering these to improve the user-

experience– Discard all items that have no description, or

title < 4, or title repeated > 100 times– Discard all items that have no description, and title < 4, or title repeated > 100 times

• Results improved but not quite sufficient• Plans: Show users “interesting” items first

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Page 8: PATHS at Digital Humanities Congress 2012

Cutting Paths through DCH

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Leeds Town Hall

Leeds City Art Gallery – Entrance

Leeds City Art Gallery - Interior

Background Information

Keywords

Search facets

Thesauri / Vocabularies

Similar Items

Page 9: PATHS at Digital Humanities Congress 2012

Similar Items

• Use Latent Dirichlet Allocation to automatically determine a set of 700 topics in the collection– Similarity calculated based on which topics an item

belongs to– Show the 25 most-similar topics to the user

• Plans:– Produce more diverse similar items– Limit based on meta-data (similar items in the from

the same source, ...)

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Page 10: PATHS at Digital Humanities Congress 2012

Keywords / Search Facets

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Page 11: PATHS at Digital Humanities Congress 2012

Thesauri / Vocabularies

• Existing meta-data is not consistent in use of thesauri / vocabularies

• Automatically map to a number of manually and automatically generated thesauri / vocabularies– LCSH, Wikipedia categories,

DBpedia ontology, Wordnet domains, Wordnet, LDA-based hierarchy, Wikipedia article hierarchy

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Page 12: PATHS at Digital Humanities Congress 2012

Background Information

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Painting of a fen landscape

A fen is one of the four main types of wetland, and is usually fed by mineral-rich surface water or groundwater

• Automatically link items to Wikipedia articles that are related

Page 13: PATHS at Digital Humanities Congress 2012

Creating & Sharing

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Leeds Town Hall

Leeds City Art Gallery – Entrance

Leeds City Art Gallery - Interior

Keywords

Search facets

Thesauri / Vocabularies

Similar Items

Add to WorkspaceCreate your own Path!

Page 14: PATHS at Digital Humanities Congress 2012

Evaluation

• Just finished the first round of full-scale evaluation (31 participants)– Thumbnails a must (bigger is better)– Contextual information appreciated– Search / Browse patterns seem to be different– Desires

• Paths that can branch• On-line help• Better integration

Digital Humanities Congress 2012, Sheffield, 6th – 8th September

Page 15: PATHS at Digital Humanities Congress 2012

Thank you for listening

[email protected]

http://www.paths-project.eu

Find out more at:

The research leading to these results has received funding from the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement no 270082. We acknowledge the contribution of all project partners involved in PATHS (see: http://www.paths-project.eu).

http://prototype.paths-project.eu

Try it out at:


Top Related