Web Mapping and Mash-ups

Jon Blower(and the participants of the "Geobrowsers"

workshop, Cambridge 2007)

jdb@mail.nerc-essc.ac.uk

Reading e-Science Centre

Environmental Systems Science Centre

University of Reading

Web 1.0: the Old World

Separate websites for

each provider of data or

functionality

Users might have to visit many websites to get their job done

Each website is a “dead end”

The Brave New World

Web Services Web Services Web Services

Different communities can combine data and functionality in their own way

Data and functionality

exposed through Web

Services

What are mash-ups and what do they do?

Combinations of data and/or functionality from more than one independent source to create a new

application or reveal new information

• Foster creativity– Doesn’t rely on one provider thinking of everything!

• Lower the cost of application development and information discovery

• Easy to use!• Lightweight Geographic Information Systems

A Search and Rescue mash-up• ESSC and BMT Cordah

Ltd• Use Google Earth as

common platform for visualizing:– Ocean forecast data

(from UK Met Office)– SAR predictions (from

BMT software)

• Can see path of drifting person and sea temperature, animated simultaneously

• Could also bring in wind data, shipping lanes etc

Web Mapping• ICEDS

– Integrated CEOS European Data Server

– UCL and ESYS plc• Displays global and

continental-scale, full-resolution geographic information

• Web interface (left)– Brings together land surface

satellite imagery, bathymetry, up-to-date clouds

• Direct access to map images and data– Through standard Web

Service interfaces

http://iceds.ge.ucl.ac.uk/

Desktop vs Web mash-ups

Web sites– E.g. Godiva2, ICEDS

• Users only need a web browser

• 2D visualization• Web site provider

controls the datasets users can see

• Can create custom functionality

Desktop apps– Google Earth, Microsoft

Virtual Earth

• "Virtual Globes"• 3D visualization• Easy to incorporate

new datasets– E.g. Via KML files

• Often can't add new functionality

Lots of overlap! Many data formats can be read by both.

More mash-up examples

Reveal spatial patterns

• Visualize species distribution alongside 3D rendering of topography– Google Earth again!

• Can see that the waxy earwort (Douinia ovata) seems to prefer valley environments

• National Biodiversity Network

• Jonathan Cooper, CEH

http://www.brc.ac.uk/GoogleEarth/NBNGoogleEarth.jsp

Visualization of scientific phenomena

• Hurricane Katrina, August 2005

• Picture left shows sea surface temperature (UK Met Office) and storm position/intensity (TRACK analysis of ECMWF data)

• Winds cause upwelling of cooler subsurface water on right-hand side of the cyclonic storm track

• (much more obvious in live system!)

Direction of missions with real-time data

• British Antarctic Survey (BAS) used Google Earth to direct 2 scientific cruises in 2006

• Multiple data streams (ship location, sea temp and salinity, air temp and pressure) streamed to Google Earth in near real time

• Combined with info about wider environment

• Enabled real-time decision-making (e.g. tracking of predators, left)King penguin track

overlain with concurrent chlorophyll and satellite

imagery

Diagnosis of models and observations

• Picture left shows comparison of NEMO model and observations for Nov 2004

• Red dots show bad model-obs fits, green dots are good fits

• Google Earth allows very efficient browsing of these large datasets

• Could read obs and model data from different locations and bring together in Google Earth or another client

• Requires some server-side functionality

Combining web mapping with virtual globes: Godiva2

http://lovejoy.nerc-essc.ac.uk:8080/ncWMS/godiva2.html

Displays dynamically-generated images of environmental data on a website

Acts as an interactive, 4-D online visualization system to allow the user to explore data and find interesting features.

Then user can view the data in Google Earth and combine with other datasets

How to get involved…

… as a user• Download Google Earth and play!• Visit some links to find out what others are doing:

– Geobrowsers workshop: www.niees.ac.uk/events/GoogleEarth – Online community: www.scispace.net/geobrowsers

• Example data available in Google Earth (KML) format:– CASIX database– Lots of US weather data– MCA hydrographic surveys– EarthNC marine charts (NOAA)– National Biodiversity Network

• Use your imagination! What would you like to see happen?– Post a message to the Geobrowsers community or send me an email!

… as a data provider• Different levels of effort depending on what you want to

achieve:– (1) Provide locations of your data in a simple format (e.g. GeoRSS)

and display on the Web using Google Maps or Worldkit.

– (2) Link the data points to your website for more information

– (3) Convert some data to Google Earth (KML) format and put on your website. Allow Google to find these files automatically (instructions available on the web).

– (4) If you have large data sets, consider generating KML dynamically through a Web Service, allowing users to choose the data they see

– (5) Put up standard Web Services (e.g. Web Map Service, Web Feature Service)

• Important to use standard data formats where possible!

Useful Web Services that provide common geographic functions

• Geocoding (Yahoo, Google)– Convert addresses to lat-lon

• Gazetteer (lots, e.g. Gaze)– Converts town, country names, IP

addresses to lat-lon– Disambiguates names

• Can envisage many more generic services

– Coordinate transformations– Data format conversion (e.g. GeoRSS

-> KML)

• Yahoo Pipes allows you to string these things together in an online workflow:

– E.g. picture right: find apartments in San Francisco near a park

– Display on a dynamic web map– Then export as KML

Discussion: the future• Increased uses of these technologies in education and

outreach• Move beyond visualizations to data manipulation

– Through creative use of Web Services and distributed computing• More scientific datasets made available through Google Earth

etc– Have started the ball rolling with "Google Ocean": in progress

• Better discovery of available data through catalogues– Google Earth can be a good discovery interface– Can EDINA play a role here?– Just being able to discover and preview data simply would be a great

step forward!• Security

– How to secure data streams whilst not losing compatibility?• Please join online community at

www.scispace.net/geobrowsers!– Share datasets, hints, experiences, applications, ask for help…

Thank you!

(the next few slides give useful references to key technologies that you

might like to investigate…)

Reference: key Web Mapping technologies

• Google Maps, Yahoo Maps, Microsoft Virtual Earth– Free but closed-source systems for dynamic web mapping

• OpenLayers:– Open-source web mapping application, interacts with standard web

services• WorldKit:

– Very simple web mapping toolkit, Flash-based• Web Map Service implementations:

– UMN Mapserver– Geoserver– ncWMS

• Web Feature Service implementations:– Geoserver– … (THAT RESTful thing that's easy to use apparently)

Reference: Virtual Globes

• Google Earth– Closed source, easy to use– Data ingested through KML

• NASA World Wind– Open source, NASA-oriented

• ArcGIS Explorer– Free, but works best when linked to ArcGIS

Server (non-free!)

• Microsoft Virtual Earth 3D• Many more!

Reference: useful data formats for mash-ups

• GeoRSS– Very simple way to geo-locate a piece of information– Can be displayed by Google Maps, OpenLayers, WorldKit, more…

• KML (Keyhole Markup Language)– Google Earth's format, but can be read by many other GIS applications, web

and desktop– Combines geo-location with Styling

• Geography Markup Language (GML)– More complex but rich– The output from a Web Feature Service

• GeoTIFF– "A picture that knows where it is"– Can't be displayed by most web browsers without a plug-in– Can also contain data

• GIF, PNG, JPEG– Don't contain geolocation information but can be displayed on the Web– The output from a Web Map Service

Reference: Open Geospatial Web Service Specifications

Web Service Purpose

Web Map Server (WMS) Serves map images (e.g. satellite images)

Web Feature Server (WFS) Serves geographical features (e.g. buoy locations, radiosonde profiles)

Web Coverage Server (WCS) Serves multidimensional raster data (e.g. numerical model output, satellite data)

Web Processing Server (WPS) Processes data

Sensor Web Enablement (SWE) – coming soon

Whole suite of standards for monitoring and managing sensor systems

Geography Markup Language (GML) underpins the Web Services