Successfully reported this slideshow.
Date: 09/06/2014
User Requirements for
Geospatial Provenance
Daniel Garijo, Andreas Harth, Yolanda Gil
Ontology Engineerin...
Problem statement
Maps can integrate many different sources
•Open Street Maps
•GeoNames
•CIA World Factbook
•Etc.
Interact...
Outline
1. Challenges
2. Assumptions
3. Types of provenance in the geospatial domain
1. Provenance of datasets and sets of...
Challenges concerning provenance
Versioning and provenance
(Map updates )
Trust based provenance Data integration and prov...
Assumptions
Simplifying the problem…
•The entities across datasets have been mapped.
•The datasets share the same data mod...
Summary
1. Challenges
2. Assumptions
3. Types of provenance in the geospatial domain
1. Provenance of datasets and sets of...
Types of provenance: Provenance of Datasets and sets of Datasets
Provenance of a map…
•Sources used to create the map
•Cre...
Types of provenance: Provenance of Objects and sets of Objects
Objects: lower granularity entities in the map
•Original da...
Types of provenance: Provenance of Properties and sets of Properties
Properties: attributes of objects in a map
•Sources o...
Other requirements related to provenance
10
Other requirements might not be straightforward to answer…
•How did a set of m...
Summary
1. Challenges
2. Assumptions
3. Types of provenance in the geospatial domain
1. Provenance of datasets and sets of...
Modeling provenance in the geospatial domain: PROV-O extension
Simple PROV-O extension to model the dataset level
12
Dataset Level Provenance: Example
13
Dataset integration approaches
There are different alternatives for updating a map
14
Object level provenance: scalability
15
Property level provenance
16
Asserted properties do not have URIs!
•New entities for describing their provenance
Source A ...
Conclusions
17
Requirements and
major challenges for
geospatial
provenance
4 main categories:
•Provenance of datasets
•Pro...
Date: 09/06/2014
User Requirements for
Geospatial Provenance
Daniel Garijo, Andreas Harth, Yolanda Gil
Ontology Engineerin...
Upcoming SlideShare
Loading in …5
×

User requirments for geospatial provenance

678 views

Published on

Presentation for the PROV analytics 2014 workshop

Published in: Education, Technology
  • Be the first to comment

  • Be the first to like this

User requirments for geospatial provenance

  1. 1. Date: 09/06/2014 User Requirements for Geospatial Provenance Daniel Garijo, Andreas Harth, Yolanda Gil Ontology Engineering Group. Universidad Politécnica de Madrid Information Sciences Institute, University of Southern California Institute AIFB, Karlsruhe Institute of Technology
  2. 2. Problem statement Maps can integrate many different sources •Open Street Maps •GeoNames •CIA World Factbook •Etc. Interaction to standarize 2
  3. 3. Outline 1. Challenges 2. Assumptions 3. Types of provenance in the geospatial domain 1. Provenance of datasets and sets of datasets 2. Provenance of objects and sets of objects 3. Provenance of properties and sets of properties 4. Other requirements related to provenance 4. Modeling geospatial provenance with PROV-O 1. Dataset level provenance • Updating a map 2. Object level provenance 3. Property level provenance 5. Summary 6. Conclusions and Future work 3
  4. 4. Challenges concerning provenance Versioning and provenance (Map updates ) Trust based provenance Data integration and provenance Crowdsourcing and provenance Granularity and provenance Aggregation and provenance 4
  5. 5. Assumptions Simplifying the problem… •The entities across datasets have been mapped. •The datasets share the same data model and vocabulary. •Each dataset contains objects with unique identifiers. •The integrated map is going to be presented to a user who is interested in using the information for some purpose. 5
  6. 6. Summary 1. Challenges 2. Assumptions 3. Types of provenance in the geospatial domain 1. Provenance of datasets and sets of datasets 2. Provenance of objects and sets of objects 3. Provenance of properties and sets of properties 4. Other requirements related to provenance 4. Modeling geospatial provenance with PROV-O 1. Dataset level provenance • Updating a map 2. Object level provenance 3. Property level provenance 5. Summary 6. Conclusions and Future work 6
  7. 7. Types of provenance: Provenance of Datasets and sets of Datasets Provenance of a map… •Sources used to create the map •Creator of the map •Creation process used (algorithms, etc.) •Recent changes of the map •Reason why the map has been updated Browsing different versions of a map… •Most recent maps •Maps from an organization •Maps created from a version of a dataset or algorithm Map release June OSM FAO GADM Integration June 7
  8. 8. Types of provenance: Provenance of Objects and sets of Objects Objects: lower granularity entities in the map •Original data source of the object •Organizations responsible for the creation of the object •Date of creation of the object •Date of insertion of the object in the map •Process of inclusion in the dataset Provenance of collections of objects… •Source of the objects of a region/area •Objects from a specific organization •Objects belonging to a type of source (e.g., crowdsourced map) •Objects introduced in the last version of the map A B C bridge stadium intersection 8
  9. 9. Types of provenance: Provenance of Properties and sets of Properties Properties: attributes of objects in a map •Sources of the property •Creator of the property •Date of the creation/update of the property •Process by which the property was added Provenance of sets of properties… •Properties of objects coming from one data source •Properties of objects belonging to a crowdsourced map •Properties of the selected objects that have the same source 9 Source A Source B Height: 20 m Length: 1 km Name: 405 Fwy overpass
  10. 10. Other requirements related to provenance 10 Other requirements might not be straightforward to answer… •How did a set of manual corrections help to improve the map? •What is new in this map? •What objects are integrated with a high confidence? •Why is an object not appearing? •General highlights of the map …but they can be addressed having provenance records
  11. 11. Summary 1. Challenges 2. Assumptions 3. Types of provenance in the geospatial domain 1. Provenance of datasets and sets of datasets 2. Provenance of objects and sets of objects 3. Provenance of properties and sets of properties 4. Other requirements related to provenance 4. Modeling geospatial provenance with PROV-O 1. Dataset level provenance • Updating a map 2. Object level provenance 3. Property level provenance 5. Summary 6. Conclusions and Future work 11
  12. 12. Modeling provenance in the geospatial domain: PROV-O extension Simple PROV-O extension to model the dataset level 12
  13. 13. Dataset Level Provenance: Example 13
  14. 14. Dataset integration approaches There are different alternatives for updating a map 14
  15. 15. Object level provenance: scalability 15
  16. 16. Property level provenance 16 Asserted properties do not have URIs! •New entities for describing their provenance Source A Source B :Bridge :height 20m :Bridge :length 1 km :Bridge :name “405 Fwy overpass” :metadata1 :metadata2 prov:wasDerivedFrom prov:wasDerivedFrom
  17. 17. Conclusions 17 Requirements and major challenges for geospatial provenance 4 main categories: •Provenance of datasets •Provenance of objects appearing in the map •Provenance of properties •Other Analogous questions are relevant for dataset/object/prop erty provenance in non-geospatial domains.
  18. 18. Date: 09/06/2014 User Requirements for Geospatial Provenance Daniel Garijo, Andreas Harth, Yolanda Gil Ontology Engineering Group. Universidad Politécnica de Madrid Information Sciences Institute, University of Southern California Institute AIFB, Karlsruhe Institute of Technology

×