Category:Datasets

From Traxel Wiki
Revision as of 18:12, 31 October 2024 by RobertBushman (talk | contribs) (→‎Links)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

To include your page here, tag it with [[Category:Datasets]].

Datasets

Fusion Data

Links

IPH

So, extremely roughly:

  • 2,000 cases with access to a gun, 200 cases without.
  • 20 per 100k with access to a gun.
  • 2 per 100k without access to a gun.
  • worst case 18 per 100k due to access to a gun. (could also be killers are more likely to have access to a gun - so maybe 10 - 15 per 100k is the realistic case)
  • 10mm abusers annually
  • 44% of households have guns
  • 4mm abusers with guns kill 2,000 people (with roughly 1500 being avoidable)
  • Each abuser has a 1:2000 chance of being a killer if they have a gun.

homicide:

  • annual homicides: 26,031 or 7.8 per 100k
  • annual gun homicides: 20,958 (81%)
  • Population 350mm
  • Each random person has at most a 1:13,461 chance of being a killer.
  • domestic abuser without a gun is 1:20,000 to kill partner
  • domestic abuser with a gun is 1:2,000 to kill partner
  • domestic abuser with a gun is 1:1,346 to kill anyone (extremely rough)
  • Which recapitulates the earlier stat: 10x as likely to kill. I think that's just how I mathed it, not an actual statistic.

Life Expectancy

Us life expectancy gap.jpg

Politics

  • Political Contemporary History Data
  • Crowd Source Data
    • Users tag data with classification tags, based on snippets from Wikipedia.
    • Browser widget lets people jump to the next needed classification tag.
    • "Needed" is defined by people putting money into refinement of the data.
      • "Needed" tips can include a desired outcome? "I'm only paying for data that confirms/contradicts my hypothesis?"
    • "Provided" is defined by ... users putting cash in to vote on data quality and an ML algorithm allocating the correctness value of each classification based on other submissions and data quality votes.
      • I think it would be unavoidable that people would tip based on whether the answer agreed with their beliefs.
      • Payment for being right (high quality) could be made as an annuity with the future having the opportunity to correct the beliefs of the past. The residual revenue stream provides and extra incentive for not being disproven.
    • Taggers get paid based on a function of "Needed" and "Provided"
  • Collaborative Filtering of Content Displayed
    • View content that matches what you've been tipping for.
    • View content based on the tip history of other identities.
    • Tip people for creating identities that have strong filter characteristics for you.

Pages in category "Datasets"

This category contains only the following page.