Drumbeat/MoJo/hackfest/berlin/projects/uberproject

Project Name: Uberproject (working title -- name suggestions welcome)

Project Lead(s): Jordan Wirfs-Brock, Neil Dawson, Saleem Khan, Chris Keller

Project Description

Big Goal for MoJo Hackfest: Working prototype of a) data surfacing b) data visualization

Key steps toward goal:

1. Input search terms, run through Little Sis API and Sunlight Foundation API, output data on people and connections (Chris)
2. Visualize people and connections with Javascript dataviz library (Jordan)
3. User experience design -- what can you do with the data once it's visualized? (Neil)
4. Understand and construct use case(s) (Saleem)

Pending needs: Help from anyone who is familiar with Javascript data visualization libraries (like the Javascript InfoVis Toolkit, d3.js, Processing.js, Highcharts etc.) -- get in touch with Jordan.

Link for more info: TK

Link for demo: TK

Link to source code if applicable: GitHub repository

Related #hacktoberfest projects:

• MetaMetaProject by The Collective Ninja Horde (For more info talk to Dan Schultz)
  
• PLESPER by Laurian Gridinoc
  
• Reveal by Tathagata Dasgupta
  

Update, end of day 1 (9/26/11):

• Codified project idea
• Made action plan
• Identified next steps

Tasks for day 2 (9/27/11):

Chris

• Built out search page to query Little Sis API. Able to query with help from Matt of BBC.
• Built out return of API query.
• Visualized query of API.

Neil

• Wireframes for (at least) the 6 main steps

Jordan

• Find a good example of a complex news story that has lots of entities (people, places, organizations) that can be visualized with network diagrams
• Pull those people/places/organizations (this can be done manually for now -- wish list is for Open Calais of MetaMetaProject to do this eventually)
• Run them through Little Sis and Sunlight Foundation (with help of Chris)
• Make a (quick and dirty) mockup of what a visualization of the context and relations
• Implement the visualization mockup using a Javascript visualization library (probably d3.js or Javascript InfoVis Tookit -- need to decide) [Note: For this can use the code Chris already started writing]

Saleem

• Identify possible information source/inputs to test the scenario
• Extract facts, data points to explore
• Identify use case(s)

Tasks for Day 3 (9/28/2011)

Chris

• Working to get Jordan tech assistance and data she needs to visualize data.
• Pulling data for Jordan.

Neil

Jordan

Saleem

• Develop use case(s) based on real-world scenarios

User Interface (Neil)

The Process

At first, I thought it worked like this:

1. Provide source
2. Extract entities
3. Manually filter
4. Draw hidden relationships (with APIs)
5. Visualise
6. Share

...but as time went on I realised that the user need not go through the process in this order, or even tackle all these steps. They could for instance simply create a visualisation by hand.

Walkthrough

The user can begin in a few ways:

• Use the 'dot' (entity) and 'line' (relationship) tools to create a network by hand
• Import from an unformatted source and extract entities and relationships to create a starting network
• Import from preformatted source to create a starting network
• Import another user's network to build on it

Thanks

• Chris and Jordan for the awesome brainstorming sessions as we sewed our ideas together
• Chris for our conversation on the relationship between entities on the timeline and the net
• Jordan for making me realise there are a multitude of ways to use this system
• Saleem for our conversation on the topic of user collaboration
• Rhiannon for her thoughts on clustering