People Suck a Naming Places
summary
Claim that people are bad at naming places and a single global ID is unrealistic, so lightweight web‑wide identifiers work better. Built calgary.com and mined Twitter activity streams, extracting explicit place IDs from Foursquare venue id 641 links and inferring others from text. Used maps APIs and open name repositories, adopted machine tags to link services, and hit practical hurdles: mechanical reuse is blocked by restrictive licenses. Demonstrates a working pipeline for mapping names to places and where it breaks.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:20If a search falls on a map...
- 0:35Places names are...
- 1:05Humans define places
- 1:36The Cathedral and the Bazaar
- 1:51One ID to Rule Them All
- 2:21TIMTOWTDIBSCINABTE There's more than one way to do it
- 2:36Name Repositories
- 2:51Geographic Indexes
- 3:06Linked Data & IDs schemas
- 3:21Activity Streams Mining
- 3:36Most Data Licenses Suck
- 3:51Be a Data Hippie
- 4:06The World is Messy
- 4:21Bootstrapping
- 4:36Places are Social Containers
related talks
04mAll The Buildings In New York
20mParody in Place: Exposing Socio-spatial Exclusions in Data-Driven Maps with Design Parody
03mInterview with Ilija Duni
24m"Are Geospatial Tools Supposed To Bring Us Closer To Place?"
30mOpening Keynote
12mInterview
18mData and the City
37mDealing With Names and Addresses Around the World