Crowdsourcing, geospa1al analysis, GIS, data quality, LBS, data uncertainty, Standards, interoperability, mobile data capture, workflow, web services, etc. Geocomputa*onal Modelling & Geospa*al Sta*s*cs (DL) Geospa*al Science & Mobile Compu*ng (SM) Geospa*al Science, SDI & OGC standards (MJ)
Didier Leibovici, Sam Meek & Mike Jackson Nottingham Geospatial Institute University of Nottingham, the UK
[email protected]! !
http://www.nottingham.ac.uk/ngi/people/didier.leibovici
mobile data capture & Quality Assurance / Conflation
QAQC: the COBWEB QAQC
.Workflow authoring tool (WoQC-AT) BPMN encoding
The QA workflow is composed of more than one QC into a workflow that may loop back /feedback to the user or to other users etc. to get additional information. (confirmatory / ensemble / linked data )
.ontology support (WoQC-O) SKOS encoding
.running WPS or app (WoQC-WPS /WoQC-app)
QAQC: the COBWEB QAQC / 7 pillars 1.LBS
mainly devoted to position accuracy
2.Cleaning
erroneous / true mistakes / intentional mistakes
3.Automatic Validation
simple checks: topology relations and attribute ranges
4.Authoritative Data Comparison
more statistical orientated to attribute values credibility as well as feature position/occurrence making the most of legacy data and metadata about quality
5.Model-Based Validation:
same as 4 but using previous crowdsourcing data or physical models or behavioural models
6. Linked Data Analysis:
looking for evidence in social media and linked data framework etc…could in fact plug into into 5. afterwards
7. Semantic Harmonisation:
conformance enrichment and harmonisation in relation to existing ontologies
See the demos on pillar and pillar 4
Zapp: Requesting information about the distant landscape
Zapp: Remote geo-tagging of data
GRASP platform CropBASE system (CFFRC)
GEOSS
10
Workflow composition / wheat model eyespot disease
Model established with Dr R. Ray et al (UNOTT)
11
OS OS & Workflow management • Open standard BPMN 2.0, XPDL (metamodel) ISO 19115 /19157 (metadata quality see also meta-propagation), OGC WPS profiling
• Open Source - Java eclipse framework development, - Together libraries (with XPDL) - pyWPS or WPS4R - OS map viewer (data map, quality map,) thumbnail picture …
12
http://c3s2i.free.fr
6 4
Density
8
10
12
density.default(x = MRSA.scan5$HSu)
2
(A) define a condi1on of sufficiency for neighbourhood sizes (B) for each spa1al feature : (B1) build a vicinity according to (A) (B2) compute a LIkOOSA sta1s1c within (B3) assign the LIkOOSA value to (C) hot -‐spot assessment of the resul1ng map
0
ScankOO method
HsSu
0.75
0.80
0.85
(A) criterion joint entropy or condi1onal entropy (B) stepwise selec1on: best ith variable according to (A) (C) Ordered set of variables spa1ally clustering
SelSOOk method CAkOO method
(A) mul1way table of coocurrences (B) data reduc1on: k-‐Correspondence Analysis (FCAk) (C) principal tensors / “best” variables
0.90
N = 5634 Bandwidth = 0.006023
0.95
