More brainstorming ideas

** information filtering pipeline for user posts: "quality detection" - discriminatory algorithm (profanity), sentiment analysis (bias detection) "classify posts into proper authority" - classification or recommender system cooperation with authorities (pscaa only one now)

next step: summarizing and aggregation public comments that are essentially the same comment

nlp - classify a report into either a air complaint, etc, from the city. how to get a department to welcome info from that site. have someone vet the posts and mail them off to the proper authority

we can further classify the emails and only send a summary to the agency

** procedurally generate user stories interfacing with google's new api or building our own model w tags for stories about "Duwamish toxic waster" "Duwamish climate change" "Duwamish wildlife" or similar semantic analysis people may need more filtering - what kind of wildlife story will they want?

maybe further semantic analysis filtering for positive stories around "Duwamish wildlife"

** recommender system: making a summary of emergent categories the map learns the topics, lists the common/new topics for this week maybe each week send a digest, or list the digest at the top of the map

maybe we can find a corpus of data about citizen complaints, and scrape that? (open 311? find it fix it?) We can train our classification system with that data. Need to be careful about publishing private data here.

Like "activity stream" brought to life