NLWIS, POC Workshop

1
Dynamic Data Access and Dynamically Generated WMS
Layers

• NLWIS, POC Workshop
• Ottawa
• July 12, 2004

2
Agenda

• Introduction to Geolinking
• System Architecture
• AAFC Case and Data Flow
• API
• Geolinking demonstration
• Lessons learned
• Conclusion recommendations
• Whats next?

3
1. Introduction

• Most data about geographic features is not stored
  in a GIS
• Clients by Postal Code
• Sales by retail outlet
• Butterflies by Ecodistrict
• Farms by Agricultural Region
• Students by School District
• It is stored in conventional databases

4
Geographic Frameworks

• Data references many geographic Frameworks
• Business
• Telephone districts
• Postal / zip codes
• Sales regions
• Political
• Provinces / states
• Counties
• Municipalities
• Electoral Districts
• Environmental
• Watersheds
• Ecodistricts / ecoregions / ecozones

5
The need for Geolinking

• There is no easy way to enable access to all this
  attribute information, so that it can be used in
  models, or mapped by a GIS
• No simple standard for how to encode geographic
  attributes or identify geographic frameworks in
  XML
• No simple way to publish attribute datasets or
  join databases across the Internet

6
How Geolinking Works
Simple XML stream
7
Geolinked Data Access Service (GDAS)
An interface specification that describes how to
provide access to
geographic attributes, in XML,

• which you can then
• Map using a WMS
• Import into a
• database
• model
• GIS
• Process against another dataset
• View in a browser
• Import into a spreadsheet

GDAS stream
8
Geolinking Service (GLS)

• An interface specification that describes how to
  join a GDAS XML stream to a datastore.

Web Map Database Model GIS
GDAS stream
Geolinking Service
9
Geolinking WMS

• A Geolinking Service that interfaces with a WMS,
  to deliver real-time mapping of attributes
  derived from any GDAS-enabled database!

GDAS stream
Ecoregion Temperature 1 18 2 24 3 26
10
GDAS/GLS requires metadata
ATTRIBUTE Annual Farm Income units Dollars
DATABASE Census of Agriculture owner Sta
tistics Canada version 2001 FRAMEWORK
Electoral Districts owner Elections
Canada version 1999
11
Geolinking requires a Spatial Framework

• The attribute data provided by a GDAS applies to
  a set of geographic features (points, lines or
  polygons).
• This set of features is called the Framework.

FRAMEWORK Electoral Districts OWNER Electio
ns Canada VERSION 1999
12
Frameworks must be identical

• For geolinking to work correctly, the GDAS and
  GLS frameworks must be identical!
• A match of Frameworks is assured if the metadata
  in the GDAS and GLS match.

FRAMEWORK ltNamegtElectoral Districts OWNER lt
Domaingtec.gc.ca VERSION ltVersiongt1999
13
The benefits of Geolinking

• Improves the reliability of data products
• because the data can be obtained directly from
  the authoritative source
• one XML stream can be used to generate multiple
  types of products
• Improves the timeliness of data products
• because products can be instantly updated
• Reduces workload for data providers,
• because the same interface can be used to deliver
  information for all common uses
• Eliminates development times for WMS layers
• because layers can be created on demand

14
What does a GDAS stream look like?

• ltFrameworkgt
• ltDatasetgt
• ltAttributegt
• lt/Attributegt
• ltRowsetgt
• ltRowgt
• ltIgt10lt/Igt
• ltVgt36712lt/Vgt
• lt/Rowgt
• ltRowgt
• ltIgt11lt/Igt
• ltVgt339164lt/Vgt
• lt/Rowgt
• ltRowgt
• ltIgt12lt/Igt
• ltVgt433604lt/Vgt

metadata
Similar to an Oracle XML dump Similar to an
HTML table
data
15
2. Status of Specifications

• GDAS version 0.9.1 is a discussion paper at OGC
• GLS version 0.9.1 is a discussion paper at OGC

16
2. Status of Implementation

• GDAS and Geolinking were developed to support the
  CGDI.
• Agriculture and Agri-Food Canada
• Statistics Canada
• Geological Survey of Canada
• Canadian Forestry Service
• Environment Canada
• GDAS version 0.8.2 and GDAS/GLS 0.9.0 has been
  implemented and is fully operational at AAFC
• GDAS version 0.9.0 has been implemented and is
  fully operational at CFS

17
3. System Architecture
Fire
18
3. AAFC Case and Data Flow
Search Registry
Get Cap
Active services
Visualize Registry
Open registry
Request
List of services
Get Cap XML
Visualize Get Cap
Get Cap list
Get Data
Produce WMS
Request
Get Data XML
Request
Visualize Data
Get Data XML
GDAS Client
19
3. AAFC Case and Data Flow conted
20
4. API
Stateless HTTP binding Using GET or POST
methods (to be implemented) Using
Key-Value-Pair encoding Using XML encoding with
SOAP stream Follow the Request and Response
roles Examples GDAS GetCapability GDAS
GetData GLS GetData http//wms1.agr.gc.ca/gdas/
gdas_geolinking.php?serviceGLSrequestGeoLinkve
rsion0.9.1dName agr.gc.cadataNameEcoregion-19
61-90ClimateNormalsaNametmean_apraTypeinterval
fVersion1fName EcoregionsgetdataHosthttp//w
ms1.agr.gc.ca/gdas/getdataAppGetData.php
21
5. Demo
GDAS/GLS on either side of the fire wall
Application for public with no restriction
(Ecostrat data) Application with authentication
measures Application inside of fire wall
(Census data) AAFC public accessible GDAS/GLS
application Open source solution e.g. Apache,
PHP and Mapserver Add new OGC specifications
i.e. GDAS/Geolinking Support multi-attribute
streaming and mapping Dynamic WMS layer
generation Utilize Distributed Computing
Platforms Leverage web mapping by using GDAS/GLS
client GO DEMO
22
6. Lessons learned

• Align projects to business requirements
• High-tech means simple and easier for users
• Seeing is believing
• Show people the prototype
• Being flexible is not always a easy thing to do
• Between Open Source and Proprietary products
• Knowing what your solution can and what cant
  solve
• Reorganize the backend database
• Web Mapping is slow back its nature
• Can one stone kill more than two birds?
• Reusable resources means cost saving

23
7. Conclusion Recommendations

• GDAS/Geolinking provides interoperability
• GDAS/Geolinking can help distribute geolinked
  data in a simple and yet efficient way.
• GDAS/Geolinking streams can be parsed and used
  for
• Statistic analysis, modeling, data integration
  and data visualization on the fly.
• GDAS/Geolinking technique has been tested and
  operational in AAFC
• GDAS/Geolinking technique should be continuously
  used
• GDAS/Geolinking should be part of the full POC

24
8. Whats next?
Solutions for efficient web mapping Multiple GDAS
streams based modeling etc. Develop generic APIs
of GDAS/GLS Setup more operational use
cases Using data access control technique Using
web service registry
25
9. Questions
Thank you very much!