Hackpads are smart collaborative documents. .

Josh Jacobson

840 days ago
Unfiled. Edited by Josh Jacobson 840 days ago
Peter H
  • Allows for a section for people to comment as relevant.
Josh J
  • I haven't used Trello all that much... but the board's name that is linked is 'small tasks'. Is there a different board for larger ones? If not, what is the key for knowing the time assigned to each task?
 
262 days ago
Unfiled. Edited by Josh Jacobson 262 days ago
Great idea - I'd be happy to put in time as a moderator/answer the questions any newcomers have, but it would have to be on a part-time basis.
Josh J
  • Group is generally seen as as poor by existing EAs. Could greatly benefit from more active moderation.
 
230 days ago
Unfiled. Edited by Josh Jacobson 230 days ago
Remmelt E
  • I think you could also make a great 3D print out of this with the DXF file I send you. This how-to was aimed at laser cutters.
Josh J
  • I bought them from Ponoko. Smart idea to post them there. That said, as an FYI, they arrived with issue #4 (a film attached to the plastic that I so far can't completely remove).
 
259 days ago
Unfiled. Edited by Josh Jacobson 259 days ago
Josh J Using the database of all 501(c)3 organizations in the US, I plan to create ratings for every charity, based on effectiveness.
 
Rationale: GiveWell and EA in general have had trouble with mass-appeal in part because tools that work on overhead ratio are more popular. These, such as Charity Navigator, have great appeal in large part because people can get a score for nearly any organization. This will create an EA alternative. The version 0 of this can be created quite easily and still be meaningful, while iterations will much improve the product.
 
Limitations: Initial scores will lack confidence as they will be based pretty much on just the organization name. Long-term scores will never approach GiveWell levels of confidence. Organizations with names with non-obvious purposes (i.e. Kiva) will have no rating until a manual review.
 
To address these limitations, clear and prominent communication on levels of confidence will be central, and manual reviews will be ongoing.
 
Plan
 
  1. Data / evaluation
  1. Use charity's names to tag them for effectiveness indicators
  1.  Examples: a disease name, a location name, a cause area name, type of action taken, etc.
  1. i.e. if organization name contains 'Malaria', it's more likely to be effective
  1. i.e. if organization name contains 'Congo', it's more likely to be effective
  1. i.e. if organization name contains 'Homeopaths', it's likely to not be effective
  1. Manually program in obvious scores, such as AMF and Homeopaths without Borders
  1. Based on the tags and number of tags assign a rating of effectiveness and a level of confidence in the rating
  1. Begin manually evaluating charities to turn this into a supervised machine learning problem, and update ratings and confidence levels for charities that have not yet been manually rated
  1. Presentation
  1. Design
  1. Front-end presentation of charity ratings
  1. Front-end index &/or search page
  1. Web programming
  1. Implement a back-end that reads from data file and updates the view
  1. Resources
  1. Needed: person or technical service for part 2b
  1. Needed: manual evaluators comfortable providing opinions on charity effectiveness based on the limited available public data (instructions and guidelines will be given)
  1. Optional: Data scientists to review / coordinate on data implementation
  1. Optional: Someone to lead the front-end presentation from the design standpoint
 
Updates/meetings
Part 1a of above plan is complete
 
  • 5 hours thus far
  • Money
  • None
 
262 days ago
Unfiled. Edited by Josh Jacobson 262 days ago
Josh J Has repeatedly been criticized as an idea that downplays the experience of poverty and may do more harm than good. Consider cessation. (Full comments on EA Forum)
 
262 days ago
Unfiled. Edited by Josh Jacobson 262 days ago
People: Josh Jacobson (and others)
Technologies: RoR, PostgreSQL
Development Stage: On hiatus, tech completely built, can be taken over by an EA as a EtG opportunity
 

Contact Support



Please check out our How-to Guide and FAQ first to see if your question is already answered! :)

If you have a feature request, please add it to this pad. Thanks!


Log in