iTudes (thanks Britt Blaser)

AKA "How's my driving?" - Social Mood Tracker

The Problem:

How do I gauge in real time a "community attitude" as it relates to a specific time, place or other point that can make it have more specificty (uniquely identifyable), or not. Tagging today can lead to too much information.

It's hot-or-not for tags in context:
"your mother" = LOVE HER
"your mother, naked" = NOT SO MUCH

Contexts:

Event, function, project, company, service, person, meme, etc.

User Classes

Critics: can create free form, heirarchical tags, then rate them - critics can store models and provide them (via link) to others to use to rank a specific tag cluster/grouping

Consumers: can pull reporting via tag combinations, weighted to create unique models per consumer. Consumer modeling could then be locked in (here's how I want to rank presentations) and used by other consumers for a true comparison.

Ranking Clusters: groupings of tags that generate statistical results based on user defined weightings of the tags - tags

Interface:

Tag Math:

• foo+bar (and)
• foo-bar (exclusion)
• foo bar (or)

Taggable and Commentable

• Comments provide additional feedback (e.g., "If the dude says "ahhhhh...." one more time...)
• Tag combinations
• Tags can be related or non-related

Linkable (Full url - http://HMD.org/rate/tag-or-topic)

Statistical (Ranking on more than one item)

• allow grouping of tags ( example : event / agenda item or topic / presenter

Feeds in from "out of box" application:

• Eventoutofabox
• http://upcoming.org/

Feeds out via:

• data generated comes in one of two forms : Instant % happiness score, and statistical graph o' happiness.
• consumer either goes to website and generates a model with all the relationships, creating a key, *or*, user passes in tags with relative weightings for ad-hoc model generation. Key from anonimity must be used to track who created the model. Javascript / Ruby / PHP / .Net SOAP calls in with consumer key and model key (which aren't necessarily tied together - ie: i can use Ryan's model key to get his statistical model) - returns two values : model name and either numerical array of happiness over 30 days (limitation!) or instant number of current happiness state.

Can't be anonymous:

• email verified for access
• key generation for access to api that provides back the model / score
• storage / reporting of the model that generates a specific ranking to lend credence to the reported ranking

Community Moderation:

• Bad ratings can be tagged as 'questionable' and when enough people ( what is enough? ) tag it, the rating gets removed
• Ratings would be tracked by person / recording the rating that they gave
• statistics can be generated not only on what the person ranked, but also on the rankor (is that the right word?)
• Overall patterns and statistics can be generated based on tag groupings ( people love / hate conferences, etc...)
• E-Bay style ranking for moderated comments

Assumptions:

• Rails based application
• MySQL data store?

Responsibilities:

• Ryan Thrash - CSS Design and over all sex appeal
• Brian Oberkirch - Process and workflow - naming the damn thing
• Kevin Marvin - Code management
• Community Rails developers and Kevin Marvin / perhaps Adam Keyes - Coding
• Alex Leverington - Generation of website, wiki, SVN, Trac

 Create your own business wiki!Support CommunityPublic, not yet premium Privacy Policy RSS