[Turing Data Story] What should I be donating to?

[kevinxufs]

Story description

Please provide a high level description of the Turing Data Story

Some charities require $11,000 to fund a wish. Others require $6000 to save a life. It is clearly important that we think about how we donate our money to have the most impact. There are currently a number of 'charity analysers' which specifically evaluate charity impact, the most prominent one being https://www.givewell.org/ who post updated lists on the most effective charities to donate to.

Somewhat surprisingly, whilst these charity analysers emphasise the importance of looking at data and real impact, they are not particularly transparent with how they reach their analysis, at least not at the level we might expect. We don't get to see the raw data they are using, nor any code on how they reach their conclusion. Part of the issue is that each analysis may be somewhat bespoke, i.e. there's no set process or metric. However this still seems unsatisfactory.

I've recently been looking for charity APIs to see if I can find data on charities myself to attempt to replicate their analysis, or to start some new analysis myself. I'm sceptical we'd quite be able to create a script to solve 'which is the best charity', but it could be interesting to see how charity themes change over time and where funding is going to. For example we might be able to track certain funding to contemporary events (e.g. The Red Cross and the recent war).

Which datasets will you be using in this Turing Data Story?
https://charitybase.uk/

This website seems to have an API for different UK charities. Interesting data fields provided include:

People, finance, themes and areas

Additional context

This is also related to effective altruism movement: https://www.effectivealtruism.org/

The movement emphasises using rational thinking to dictate our donation, thus it would be great to be able to genuinely evaluate some data and come to our own informed conclusions.

Ethical guideline

Ideally a Turing Data Story has these properties and follows the 5 safes framework.

• The analysis you produce is openly available and reproducible.
• Any data used are open and have an explicit licence, provenance and attribution.
• Any data used are not personal data (i.e. the data is anonymous or anonymised).
• Any linkage of datasets in your data story does not lead to an increased risk of the personal identification of individuals.
• The Story must be truthful and clear about any limitations of analysis (and potential biases in data).
• The Story will not lead to negative social outcomes, such as (but not limited to) increasing discrimination or injustice.

Current status

• Scoping of the story
• Write story outline
• Add material to the story notebook
• Combine materials into a readable story
• Proofread
• Request reviews
• Address reviews
• Merge to master branch.

Updates

[kevinxufs]

I've sent a message to GiveWell now:

Hi,

I'm part of an open source project in the Alan Turing Institute (UK's national data science research institute) called Turing Data Stories. The aim of our project is to write transparent, reproducible and educational data analysis on different topics. We'd be interested in a chat to discuss a potential collaboration, you can see our Github here
https://github.com/alan-turing-institute/TuringDataStories.

The main reason I'm reaching out is because I'm a supporter of effective altruism and of the Givewell principles. Primarily, I really value having transparent and quantifiable analysis in making these decisions. When looking at the GiveWell website I noticed that there was transparency in terms of where the money was donated to, but I think it would be interesting to see the GiveWell methodology, as well as understand your data sources as I think this will increase overall trust.

Turing Data Stories is a not for profit group composed of volunteers interested in using data for good. This includes data scientists who are keen to analyse interesting data as well as story writing / editors who help formulate an interesting narrative around this data, as well as ensure our content is educational. A potential collaboration would not require any technical resources from GiveWell, instead just relevant data and someone we could check in with.

It would be great to have a call to discuss this.

Thanks,

Kevin

[kevinxufs]

Just to note that we had a meeting with the GiveWell outreach team. They mentioned they were quite stretched so would be difficult to invest time acting as a reviewer or providing guidance on their methodology.

They posted a few links to look at, for example their cost effectiveness analysis: https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models

Given this, I'll take a look through their website and see if there's any particular information that we can do quantitative analysis on. If I find anything we can reach out back to them with a proposal.