13 Jan 2016

HappiApp or AdKiller?

  1. Researchers often publish articles whose "significant" results cannot be reproduced or acted on. The results sometimes reflect the influence of missing variables or ideosyncratic data.
  2. Governments want to ensure that people are as happy as possible (by their own definition), if only to reduce spending on crime or depression or increase productivity.
  3. Most people would agree with me, that the bias in advertising taxes social efficiency.
What do these three statements share in common? They all point to an idea of mine (and perhaps others), namely, to make an app that tracks and improves happiness.

Feel free to implement this idea if you're that type (my "source fee" is one beer.)

How does the app work? I've copied the proposal below, but here's a step-by-step guide:
  1. Ask a bunch of researchers to give you a list of "activities that make people happy"
  2. Add those to the app as a list to choose, along with "[not on this list]"
  3. Release the app.
  4. It will ask people a few times per day what they are doing and how happy they are
  5. These results will be stored for the user and also -- via machine learning -- generalized
  6. They will also falsify earlier reseach that may have found spurious correlations with happiness
  7. Updates to the app will update questions and the algorithm, to "improve" but not "maximize" moods
  8. The biggest target is the needy fear of advertisements and other propaganda.


Working name: HappiApp
Audience: People interested in tracking their mood and activities
Cost: Free (no advertising)
Interface: Users supply basic demographic and geographic data
Process: Users choose 1-20 “prompts” per day. On each prompt, the app asks how happy they are (1-5 scale) and what they are doing (drop down)
  1. Pairs (happiness, activity) are matched against academic literature to see if existing correlations are (not) falsified.
  2. “Learning algorithm” can suggest “happier” activities based on (a) user history and/or (b) activities pursued by similar users (in terms of other patterns)
Funding sources:
  • Research organizations interested in data (correlations)
  • National organizations interested in results (happier people)
  • Fremium model that includes more information on options and/or greater detail to data
Privacy concerns: None. Opt-in model. No personal information (name, address, credit card) collected.
Motivation: Give people more feedback on how to improve their lives from source outside advertising/commerical industry.
Competition: Existing Apps are more likely to provide “dumb” check list of things to do, without learning and/or cross check with other users.

No comments:

Post a Comment

Note: only a member of this blog may post a comment.