Where 2.0: From Data Chaos to Actionable Intelligence
Sean Gorman
[First slide was missing]
Three trends: Geoweb, handling large datasets, emerging semantic web
Story began as: Trying to GeoHack Without getting arrested
Sean started as a geographer but didn’t want to be a geographer really, looking at conceptualizing the internet, what the router graph looked like, etc. etc.
Big Data Sets -> Algorithmic Analysis -> Map Logical Results to physical realities
Made for some pretty interesting maps - “How to take down NYSE?”
Men in black suits turn up.
IN-Q-TEL came to the rescue, big history of taking ideas to market from the academic and start-up culure. Specifically in the geo-web. Have taken others, MetaCarta, keyhold, Last Software.
Tip of the iceberg that the role that the government has played to take these to market before there’s a big market, like GPS, satellite imagery.
Geography on the web became mainstream as we’d been playing around with geography of the web.
We’d always been focused on really large datasets. Stuff coming out was based on small
Geocommons - “Crowdsourcing large structured data sets with quantitative capabilities”
“Didn’t we try this last year”
“What happens when your database reaches 1,683,185,246 features”
Database goes a bit kaput.
Why so fast? Mainly it’s been on the long tail, we’ve been working on the short tail. 95% of the data
Data Normalization -> Tables get big -> Optimization -> Data Ingress -> Repeat
Everything helped but nothing really solved the problem.
Continue to fight
or
Build a lightweight object database
Can fit well over 1 billion features into ~16 Gigs of storage
Launching “Finder!”
Can search and can add your own data.
Link GIS and the real world.
[Demo]
Describe data, tag it, URLs, metadata URL for proper GIS marker, contact information…
Data was mined as it was uploaded. Can see statistics based on the data after uploading.
Can pull data out into KML, GML. Can view in Google Earth, MS Virtual Earth.
Links to metadata, FTGC[?], iso standard, micro formats.
Showing a KML file in virtual earth.
Census demographic information can be overlaid with sales information entered as part of the demo.
Can download the demographic information, pull it into google earth, can also pull it into GIS workflow.
[Demo showing polygon areas highlighted in colours on google earth with masses of information in an info window]
Expanding to: Maker! - ? Atlas! - share maps around stories and collaborate
Talking to other silos of data about trying to federate our data and interconnect it. Where the cloud and the data can be owned by everyone.
- Bringing a new class of content/data to the web
- Intelligently….
- Enable the content to answer meaningful questions for users
From Data Chaos to Actionable Intelligence
Technorati tags: geocommons, where, where2.0, where2008
Comments
— Kyle
— john