2012.09.12
Stanford University
Historypin / We Are What We Do
Industry Advisory Panel Meeting
We met today to discuss 3 projects running under a Mellon grant at the Stanford University Center for Spatial and Textual Analysis (CESTA).
Themes for the day for me were:
- Quality over quantity helps build attention, and thus, community
- Look for levels of granularity on materials to be crowdsourced. There’s a spectrum of work made up of tasks of increasing resolution.
- Make end states for things so people see an end. Consider different end states for individuals (10 things!), collections (every page of an atlas, a chapter in a novel) etc.
- Become friends with the Chinese Californian History community for the Railroads project.
Year of the Bay
- Content gets more meaningful with levels of specificity.
- Specific incentives around tasks or areas
- Physical events that facilitate data collection will be useful, like the Alma voyage. Consider taking digitization tools onboard.
- Several libraries, archives, museums are doing exhibitions during the Year of the Bay
- "Curating the Bay: Crowdsourcing a New History"
- Field Papers tie-in? Perhaps a map could be handed out at events.
- Older audiences appreciate regular, predictable experience vs flash mobs turning up at an instant.
- Anna (Stanford) can help with "low-level" digitization of materials. Expecting institutions to provide digital media.
- Gathering existing information
- Encouraging personal contribution
- Minimal resources
- Local champions? "Whales"? Community gatekeepers?
- Show levels of competition between museums?
- Historypin "community engagement strategy" - would love to see what that means for each project, fleshed out and scheduled.
- Mariposa/Hunter's Point Yacht Club could be a site for landing the Alma at Hunters Point
- "least digital"
- Want to make some tasks harder than the Maptcha interaction pace to illicit more detailed and personal responses.
- Mysteries, tasks
- School or student-based activities? Light curriculum adoptions? Geocaching etc?
- How would we measure results? Quality?
- Make sure there's a pay-off. Don't just ask people to "put beans in a bucket" - design a way that the system gives people something cooked at the end. Challenge/reward.
- 5 participants is enough to test with to know if something's working.
500 Novels
- Pride & Prejudice social network? Both readers and within the book itself.
- 19th century collection of 250 UK and 250 US classic novels. Copyright non-issue.
- Social network, plot of novels, each with have a general genre prescribed right now.
- Tasks
- "Who is speaking what to whom?"
- Identify which person is being referred to by each pronoun
- What kinds of actions connect people? What does she *do* to Darcy?
- Places mentioned in what context
- More public, community presence, show others’ progress or presence
- Ways to contribute? Join a novel; stalk a character
- Ask people to select a theme to help choose where to start - "Like romance novels? Try this!"
- "Chunk size"; quotes could be tweets
- Could tweet progress…?
- Lots of old books were serialized, perhaps that's a way
- Get a task/passage a day?
- Twitter game? Tweet sentences + #whoisthisabout #isthereaplaceinthis ?
- Important to describe goals around value of research
- Make more clear the research goals of the project
- Build incentive for people to help
- “Bring a book back to life!”
- Contemporary novels have significant online presence. Could that community be approached?
- What's the "reveal" of this project - genre patterns? gender relations? etc
- Field Papers approach? Give people a page or two…
- Visualize the book's social network as it's fleshed out. Daily/Hourly snapshots
- First pass computational - "all the 'she' occurrences"
- Second pass human - "help the dumb computers"
- Random sample of excerpts from Pride & Prejudice? How far could that get us?
- "We want to know 'Liz' refers to Elizabeth"
- Start with first page, then last page, then random page?
Railroads
- “Shaping the West” Spatial History lab project; 5 years old
- Many cities are located based on how far trains could get
- There is a skeleton of geo-data to build on
- How do we go out into the community about the relationship between railroads and the surrounding environment?
- Rail buffs/knowledge of rail are just one dimension
- How do we learn about changing environment?
- Pilot project: “Railroad Repeats” photography, places shot by Alfred Hart along Central Pacific Railroad
- Can rail buffs help us push out into the landscape too? Locations of stations, local environs etc
- 2 sets of rail lines digitized enough to be useful
- Much of the San Joaquin Delta will be developed in 2 ways: pre- and post-railroad. How was drainage affected? Who owned the land?
- Central Pacific: How does development proceed along a railway line? How quickly does the material world change?
- Trying to locate all supporting material in space & time then make it useful in some analytical form
- Post photos that we know are around a railway line. Tap into “buff knowledge”
- Buff Community has very fine-grained knowledge.
- Historypin building geotagging interface; then looking to add more social stuff on top
- HP wants to uncover “boxes under the bed” stories; rich personal material.
- Easy to get material out of people who owned railroads; harder to find from workers
- Post office as marker of community development (and schools)
- How does population change in proximity to a railway?
- Chinese Historical Society? http://www.chsa.org/, http://cprr.org/Museum/Chinese_Syllabus.html, http://podpdf.com/ebook/chinese-labor-railroad-pdf.html - poor records, “John Chinaman” etc
- Many original surveys of the US were fraudulent - need a crowd with significant amount of skill and self-checking - land history is most difficult
- Any site is the site of multiple stories. Interest not in correctness, but range. Dispute is interesting.
- Railroads get you back to 1860s.
- Library of Congress has a little bit of stuff: http://www.loc.gov/pictures/search/?q=chinese%20california%20railroad
- http://www.flickr.com/search/?w=commons&q=railroad
No comments:
Post a Comment