collaborative semantic geospatial applications data sources to make mashups correct, complete,...

79
Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite LLP www.giswebsite.com 14 May, 2008

Upload: kevin-mcdowell

Post on 20-Jan-2016

224 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Collaborative Semantic Geospatial Applications

Data Sources to Make Mashups Correct, Complete, Relevant and Revisited

Jonathan W. Lowe

Owner, Giswebsite LLP

www.giswebsite.com

14 May, 2008

Page 2: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Jonathan's shelves of the Lowe Family bookshelf

Page 3: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

“Will O'Reilly stop publishing new technical books when they run out of colophon animals for the covers?”

Page 4: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

O'Reilly brings fiction to a formerly technical audience

Brady Forrest

Hackerin the Rye

Handling Angst with Reg Exp

Brady ForrestBrady Forrest & Leo Tolstoy

Epic Programming Tips Magical Realism Power Tools

Perl in the Time of Cholera

War & Peace Cookbook

Page 5: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

The Goal:• People repeatedly visit your web application because its results are relevant to their specific individual needs.

How to deliver more relevant results?• Use Open, Structured Data – Communities contribute content.• Use Faceted Browsing – personalized spatial analysis.

A Whirlwind Tour of Metaweb's Freebase:• Free, open and typed data• Community of public participation• Structured data relationships• Web-based data access API• Metaweb Technologies – the company

What's this presentation about?

Page 6: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

A growing supply of community supported, structured data...

Page 7: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

OpenStreetMap: user-generated, folksonomically structured, queryable

Map tiles? Great.Source data? Better still!

Page 8: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

FortiusOne's GeoCommons: various data sources, hybrid folksonomic structure,

downloadable in several formats

Deep spatial resources for analysis and aggregation

Page 9: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

GoogleBase: user-generated, folksonomically structured, queryable

GoogleBase is spatial?Yes: GoogleBase's Business locations data enables applications such as Walk Score.

Page 10: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Domain-specific collaborative and/or structured data stores

Page 11: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Metaweb's Freebase: free, open, semantically structured, queryable

Freebase's infrastructure built and managed by Metaweb Technologies, Inc.

• $15mil VC (April, 2005)

• $42mil VC (Jan, 2008)

• Staff of 65 people

• 3 year runway

Open data strategy guided by long-term vision of co-founder, Danny Hillis.

Page 12: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Freebase's Current Domains

Freebase holds spatial data, but is not exclusively spatial

Page 13: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Freebase's 262,000 Point Locations

Page 14: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

• Points• Lines• Polygons • Non-spatial properties• Related properties

Accessing Freebase's Data

• GeoJSON• KML• ESRI Shapefile

SPATIALOUTPUT

FORMATS

Page 15: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Application Development Tools: Metacarta's OpenLayers

Freebase data: all Berkeley schools, all grade levels, public and private

Page 16: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Application Development Tools: Google Maps & Javascript

Freebase data: public elementary schools and their catchment zones

Page 17: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Freebase data: Berkeley liquefaction zones, landslide zones, earthquake faults and public/private high schools

Application Development Tools: UMN Mapserver & Perl Mapscript

Page 18: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is semantically structured data?

• Strongly typed data

• Hardwired data relationships

Page 19: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

1) Search the web for a person, place or thing.

Page 20: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

1) Search the web for a person, place or thing.

2) From the results, extract the descriptive text, but filter out the search terms and links.

Page 21: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

3) Generate a tag cloud from the extracted text.

(This example useshttp://tagcloud.oclc.org/tagcloud/TagCloudDemo)

1) Search the web for a person, place or thing.

2) From the results, extract the descriptive text, but filter out the search terms and links.

Page 22: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

3) Generate a tag cloud from the extracted text.

(This example useshttp://tagcloud.oclc.org/tagcloud/TagCloudDemo)

1) Search the web for a person, place or thing.

2) From the results, extract the descriptive text, but filter out the search terms and links.

4) People unaware of the original person, place or thing guess what it was using only the tag cloud.

Page 23: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 24: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 25: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 26: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 27: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 28: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 29: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is

Chez Panisse

?

• Gourmet restaurant serving California cuisine• Co-founded by Alice Waters• Located in North Berkeley, California• Recipies of popular dishes are revealed in cookbooks by Alice Waters and Paul Bertolli

Page 30: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Our brains are sophisticated semantic processors – what did yours just do?

• Type or classify data

• Form relationships between data

Page 31: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... alice is a name waters is a name alice waters is a person

Page 32: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

I think... a co-founder is a person

Chez Panisse

Page 33: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... alice waters is a person a co-founder is a person

Page 34: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... alice waters is a person a co-founder is a person

Hey! Both terms are about people...

Maybe: alice waters is a co-founder?

Page 35: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... a restaurant is a business

Page 36: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... co-founders start businesses

Page 37: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... a restaurant is a business co-founders start businesses

Page 38: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... a restaurant is a business co-founders start businesses

Holmes! Both terms are about businesses...

Yes, Watson, maybe... restaurants have co-founders?

Page 39: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... alice waters is a co-founder restaurants have co-founders

Page 40: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... alice waters is a co-founder restaurants have co-founders

By George! Both terms are about co-founders...

Is alice waters the co-founder of a restaurant?

Page 41: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

I think... california, usa is a place bay area is a place North berkeley is a place 94709 is a place gourmet ghetto is a place 1517 shattuck ave is a place

Chez Panisse

Page 42: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... restaurants are located in places

Page 43: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Zowie, Batman! These terms are all about places...

What restaurants are located in berkeley, the gourmet ghetto, 94709 and all these other places?

Chez PanisseI think... restaurants are located in places berkeley is a place gourmet ghetto is a place 94709 is a place ...etc...

Page 44: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Chez PanisseI think... restaurants have co-founders restaurants are located in placesand... alice waters is a co-founder berkeley is a place

Page 45: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What restaurant is located in berkeley and was co-founded by alice waters?

Chez PanisseI think... restaurants have co-founders restaurants are located in placesand... alice waters is a co-founder berkeley is a place

Page 46: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 47: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 48: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is Chez Panisse?

Page 49: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is Chez Panisse?

Who is Alice Waters?

Page 50: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is Chez Panisse?

Who is Alice Waters?

Page 51: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is Chez Panisse?

Who is Alice Waters?

Where is Chez Panisse?

Page 52: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is Chez Panisse?

Who is Alice Waters?

Where is Chez Panisse?

Where is Alice Waters?

Page 53: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is Chez Panisse?

Who is Alice Waters?

Where is Chez Panisse?

Where is Alice Waters?

Page 54: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

What is Chez Panisse?

Who is Alice Waters?

Where is Chez Panisse?

What is Berkeley, California?

Page 55: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

A Specific Individual Need:

I just graduated from Stanford Graduate School of Business and want to work for a venture-funded company in Silicon Valley (California). Maybe my connection to Stanford can help me get an interview...

The resulting query:

Find venture-funded companies in the City of San Jose, California, that employ one or more people who are alumni of the Stanford Graduate School of Business, and, while you're at it, list those employee's names.

Freebase's results for San Jose, California:

Relevance via Freebase Query that joins people, businesses and location

PayPal

• Roelof Botha

eBay

• Ben Hanna• Jeffrey Jordan

Cisco Systems

• Barry Eggers• John Turk• Greg Sands

Page 56: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

People – Business – Location

Using Freebase's hardwired data relationshipsto deliver relevant results

• Relevance in Search Results

• Relevance via Faceted Browsing

Page 57: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 58: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 59: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 60: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 61: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 62: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 63: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 64: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 65: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 66: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 67: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 68: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 69: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Relevance via Faceted Browsing

Page 70: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 71: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 72: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 73: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 74: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 75: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 76: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 77: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite
Page 78: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Relevance via Faceted Browsing

Tools such as Exhibit from MIT's Simile Project are made to visualize semantic data such as Freebase's.

Users can filter by any different property, personalizing the resulting map relevant to their specific needs.

Because the data has a semantic structure, Exhibit can automatically create new filters when users add new properties to the data.

Page 79: Collaborative Semantic Geospatial Applications Data Sources to Make Mashups Correct, Complete, Relevant and Revisited Jonathan W. Lowe Owner, Giswebsite

Useable data• The data is free and open.• The data's license does not limit its use.

Community data• All users can contribute new data.• All users can modify existing data.

Organized data• The data's explicit structure conveys its meaning.• The data's structure unifies the whole collection.

Accessible data• The data source has a web-based query interface.• The data source tracks and publishes all changes.

A forseeable future• The company running the service has funding.• The service is built on scalable technology.

What makes a data source good for relevant web mapping applications?

Jonathan W Lowewww.giswebsite.com

Metaweb's Freebasewww.freebase.com

Thank you!