Mar 2009: create database, MAP API

Apr 2009: write stylesheets, test scripts

May 2009: live test

Requirements

There are various types of exchange ‘ties’ or 'relationships' (symmetrical/asymmetrical) that exist between individual authors and stories in the Urban Collaborative:
- NARRATES
- RESPONDS
- ANNOTATES
- REVISES
- JOINS

A visualization tool should optimally be able to extract this information from the story data. We can indicate whether a tie is symmetrical (if the connection yields information applicable to both griots who establish the link) or assymmetrical (if the connection yields information that is only applicable to one of the two griots; in that case, the relation is defined in the direction of application).

Most of these relational attributes will be extracted automatically from the database. Additional graph predicates will be necessary for family and network relationships (PARENT-OF, SIBLING-OF, MARRIED-TO, CHILD-OF, RELATED-TO).

Here is an excerpt from the XSD schema, showing the model of a story.

A cleaner, more detailed version of the database schema is shown here.

I will give some examples of what the information in this graph means, and, how it may be used in the navigation of the stories:

CREATES(x,y): “Person x creates Story y; this is a symmetrical tie” - this will yield information about authorship ties and collaborative authorship of shared stories

RESPONDS(x,y): “Person x responds to Story y; this is an assymmetrical tie;” an inversion of this tie should automatically yield the information “story y is inspiration for x” - this will yield information about the revision history of narratives

ANNOTATES(x,y): “Person x annotates Story y as one mentioned in the audio text”; symmetrical - this will yield information about references and shared histories

JOINS(x,y): “Person x joins group for Story y” - this will yield information about social networks, trust and cooperation

VIBES(x,y): “Person x vibes with person y” - this will yield information about shared interests, as shown by similarities in keywords, social networks or story genre

Here are some initial thoughts for 3 basic navigation views: Page View - list stories by genre, displaying seal, author, title, group and role) Tree Network - graph stories in a network, revealing the following information in the nodes and arcs: Mentor (count), Authors, Group (site), Stories (genre), Place Itemized View - retrieve individual story based on a specific advanced database query

Our basic data model is:

Resource entries consist of persons, communities, story, transactions and comments implemented in a RESTful architecture via Uniform Resource Identifiers. We create a URI space within which static and dynamic content can coexist and be accessed through simple naming rules:

http://urbangriot.net/person/1
http://urbangriot.net/transaction/11
http://urbangriot.net/community/oasis

Authors, Swap, Group and Media content are stored in one main XML file. Doing an XSLT transformation, the XML is transformed to a format that can be easily displayed by the Google Maps API, with each author element translated into a marker element. The audio file resource is embedded into the HTML using an embed tag. A sample XML file for the visualization of a group of authors on the social network is shown here.

The swap application uses the Google Maps API and makes requests to the server using the GXmlhttp object. The request accesses the authors.xml file hosted on the server using asynchronous processing. After obtaining the array of markers via the GetElementsByTagName method and looping through it, the markers are over-layed onto the map and displayed.

Visit the prototype application to see the Urban Griots Collaborative in action.

Process

The Word. Sound. Life. Social Network for Media Learning consists of approximately 300 members grouped into 5 primary groups (learning arenas). Through a bi-weekly story swap process, selected youth advocates publish audio files. In order to provide access to these oral histories for non-members, we employ a map mash-up that combines a scraping tool, an XML-based blog using Ruby on Rails and the Google Maps API.

The blog will identify correspondences/similarities between authors and stories, including activity based on social roles or geographical location. The user may manipulate the blog content to view the results and explore different facets of the corpus. The blog content is scraped from a dedicated group page on the Ning social network and stored as XML in the database with the link to the audio for each author as the primary resource. Additional metadata is also captured and provides meaningful channels to drive the visualization engine.

Below a visual representation of the process:

Download the script for our needs assessment usability test (dates: 2/25/09 - 3/6/09).

Download the terms of use for our functioning prototype.