Hacker News Reader

One possible improvement is to provide a way for the community at large to discuss/criticize possible problems with the data as well as for Our World in Data to respond to criticisms from the public.

Recently, Bill Gates issued a Tweet demonstrating excitement for progress in diminishing global poverty, citing Our World in Data: https://twitter.com/BillGates/status/1086662632587907072

This received substantial push back, with claims that this data is a reckless extrapolation of sketchy sources, such as this article by Jason Hickel: https://www.theguardian.com/commentisfree/2019/jan/29/bill-g...

Is there anything Our World in Data can do to better facilitate this discussion, note that it is happening, or respond to criticism?

A community discussion forum to discuss research would be great. I'd be interested in hearing what you would suggest? We have discussed whether something like StackOverflow, but for global development, would be possible. I think we are too small for it though. If we'd build a platform we'd need a big group of engaged contributors and we don't think we are big enough for that yet.

I read the newspaper article and have been in a lot of contact with the author, we agree on many points, especially that higher poverty lines are needed to track what is happening and that it is surely not enough to look at extreme poverty. But low poverty lines are needed so that we see what is happening to world's poorest. One of the biggest failures of development over the last decades is that incomes of the very poorest on the planet have not risen: https://voxeu.org/article/assessing-progress-poorest-new-evi... This is not widely known because the extreme poverty line is not low enough to focus on what happens to the very poorest. So overall I think it is very important that we keep track of what happens relative to different poverty lines (and we do https://ourworldindata.org/poverty-at-higher-poverty-lines) On the particular point you emphasize, that historians have a poor understanding of poverty and prosperity in the past I do not agree that it is a 'reckless extrapolation of sketchy sources'. I think this is overstating the existing uncertainties and it's not a fair description of the careful and tedious work that historians do. And we did respond to it here https://ourworldindata.org/extreme-history-methods and go in some detail of this historical work. We understand our job as providing access to good, existing research and so our job is to understand the research that is out there and bring it together on the web to make it possible for everyone to understand it, and access it. A very thorough book on what we know about poverty, including the historical decline of the share living on less than an extremely low poverty threshold, is Martin Ravallion's 'The Economics of Poverty: History, Measurement, and Policy'. Let me know what you think!

We have been trying to engage the international development community in working with crowdsourcing of content in a wiki around water issues, https://akvopedia.org It hasn’t been very successful.

I have seen efforts around using a StackOverflow model to get engagement, but it hasn’t been successful.

The way we ended up getting content varied. We convinced those that had content to allow us to republish it in wiki suitable format. We hired an editor to do a lot of the work. We built in content publishing in the Akvopedia as a way to publish results from programmes.

It is interesting how reluctant a group of professionals can be at contributing, despite it being an appreciated source. We hear from UN organisations HQ in New York that when they get new staff they send them to the Akvopedia to read up on the subject. We know thousands of people read the content, from all over the world. I meet people at conferences that used the resource.

We specifically licensed the content to be portable to the Wikipedia, but the Wikipedia editors we have encountered are not very helpful when trying to move content over or linking content. The exception has been some content we have managed to get onto Wikiversity. But even there we met some resistance.

I think it is really important that we try different methods to spread this type of information. Something will stick at some point I am sure. Keep up the good work.

Thanks for sharing your experiences on this. I hadn't heard of Akvopedia before, so it's new to me.

Agree that finding the right solution that sticks: is read by many, consistently kept up-to-date and gets the level so detail correct/unbiased is difficult to do.

We're trying our best to do our version of this work. It's reaching some people, but we can always do more/better. It's really helpful to hear experiences/responses from elsewhere which we can learn what works and doesn't.

Appreciate the response. I wish I had some great ideas here, but I don't. Maybe allow a discussion area, but only from those willing to go through some kind of accreditation process?

The blog post that you link (https://ourworldindata.org/extreme-history-methods) is a great follow-up. But it looks like there is no connection to it from the source page (https://ourworldindata.org/a-history-of-global-living-condit...), nor the source charts. Perhaps simply figuring out a way to connect your own commentary to the data it is commenting on would be a place to start.

It's funny that you make these suggestions, because they're all things we're grappling with/considering at the moment.

Until now, most of our work has been supply-driven: we write and work on what we think people want to know rather than actually asking them. This has probably been a mistake. We're really changing the way we work to be more demand-driven. This means possibly having some kind of area where people can ask questions, have discussions etc. It's definitely something we're considering.

And you're also right that we need to work on the connectivity between aspects of the site. We're in the middle of trying different designs and formats to try to nail down the best way to do this.

I'm on mobile, so I can't provide you the full run-down. In short, I feel like the main problem is an understanding of "development" that may agree with popular conceptions, the UN, and the concept used by economists. However, this doesn't incorporate the thoughtful critiques of critical scholars that show the limits of such a narrow concept centered around economic development. There's a lot on how the cultural, social, political, and ecological dimensions are neglected, how it's operationalizations cover only a tiny sector of the economy (ignoring nature's and subsistence economy which are reduced whenever the financial economy is expanded), or how it reinforces Western and North-South power relations.

There's a lot of literature you might use to read up on this topic, especially critiques of the concept of sustainable topic. Of the top of my head, I can only think of "Ecofeminism" by Vandana Shiva & Maria Mies. While I personally disagree with the conclusions/recommendations of the book, this revolutionary book provides a thoughtful critique of development.

I really doubt that people working on development in 2019 haven't even heard of critical development studies, especially Shiva who's been making these arguments for like 40 years.

Like, twenty years ago, a basic education on the topic would make the notion of "how the cultural, social, political, and ecological dimensions are neglected" a primary subject.

Every "popular" economist and UN official has doubtless engaged with these ideas. The fact that their work doesn't revolve around them or immediately agree with every component of them doesn't mean the development community is following narrow ideologies from 60 years ago.

A subreddit would be great, no minimum size to open

I didn't think of that. Sounds like a good idea!

Or a Kialo argument tree, in case you want to facilitate, capture and visualize thoughtful reasoning in a multi layered argument tree.

Here is one I participated in about G censoring in CN: https://www.kialo.com/18304/

I had been working on a pedagogical approach to social impact learning for many years now.

http://www.developfy.com became AmplifyAI.

The Guardian article by Hickel is misinforming readers.

Hickel just wants attention for himself. Nothing else.

He does not care about the steps the world's poorest have made. Yes, small steps, but it's important and it's awful to spread misinformation like he does.

(Also, for The Guardian he chose the headline 'Bill Gates says poverty is decreasing. He couldn’t be more wrong' and elsewhere Hickel is acknowledging that this is not true and is contradicting himself: https://www.cgdev.org/blog/12-things-we-can-agree-about-glob...)

I think a more charitable interpretation of Hickel's article is that all people experience confirmation bias and much of the data presented on OurWorldInData is inconsistent with the Hickel's political beliefs.

It is not just Hickel that experiences this. Many people have a world view and political view that assumes that the current world is in terrible shape and the prevailing political establishment is failing all of us. When people with these entrenched views encounter the trends cataloged on OurWorldInData, a great deal of motivated reasoning occurs in order to discredit the trends.

"Facts can be a powerful weapon against fear, a gloomy worldview, learned helplessness". This is so true.

However I also notice a frequent and not so helpful use of such data, namely as a way of countering people's legitimate complains about their current state. It is especially concerning when the positive trends are touted as consequences of certain public policies, even when the correlation is questionable. There are heated online debates between the Bill Gates/Pinker/Factfulness camp and the 'left' on the role of the free market vs a strong state in shaping progress during the past 50 years.

How can you avoid any sort of bias creeping in the content of your site?

You're absolutely correct that we need to be very careful to ensure bias does not creep into our work.

We take a lot of time to ensure that our work is as objective as possible. Our core activity is to present the data on a given topic clearly. Where possible we try to provide a summary of what the academic research says on this topic: why a given change might have happened, why it happened in a certain way, if we know anything about direct causations. However, causation is particularly difficult: especially when there are so many dimensions interacting (e.g. poverty impacts on fertility, child mortality, education, health outcomes, and their knock-on or cyclical impacts) and a range of interacting external factors (policy decisions, trade liberalisation etc.). Unless there is solid research to support an explanation of change, we don't include it, and stick what has changed rather than why.

Maintaining a very objective, somewhat central position is important for our work. We're happy that both people on the right and left use and trust our work: if we can at least first agree on the facts and data, then we are all in a better position to then discuss why given changes have happened. Media Bias Check (https://mediabiasfactcheck.com/our-world-in-data/) puts us very slightly to the left, but we're very happy that 80% of the user responses rate us 'least biased' i.e. straight down the middle.

In terms of dismissing legitimate concerns about the current state: we think it's really important that acknowledging progress in many areas does not translate into an acceptance of the current state of the world. As we write here (https://ourworldindata.org/much-better-awful-can-be-better) although we have seen improvements in many ways, the state of the world is still unacceptable. And in some cases, things are sliding backwards (e.g. https://ourworldindata.org/homelessness-rise-england). We need to do much more. We hope that in providing a historical outlook on how things have changed provides insights and understanding of how we make the world a better place moving forward.

Have any recommendations for good pieces related to this debate you mention?

I'm sure there are lots of heated online debates, but any good ones in mind?

The top comment in the thread has the most recent debate I know of.

This is fantastic. I've been waiting for data driven policy initiatives, and I'm sure this is a great resource for that. Have you considered implementing a system like Kaggle Kernels that allow people to submit further models and analysis of the data?

That's great to hear - we're working hard to make sure our work also reaches many of the key people influencing policy initiatives and decisions.

That's a great suggestion - we hadn't actually thought of that in the context of our work. But it makes sense given we have such an abundance of open-source data.

What Kaggle Kernels do is awesome. It'd be cool to integrate this concept into our work at some point. Thanks for the suggestion!

Absolutely, it would also allow people to have a conversation about the data through various kernels and threads. As another comment said, there was some recent controversy over some of the reports, a Kernel system would enable an environment where people can have a data driven debate about what the data really implies. Thanks for your work and I'm really looking forward to seeing how it evolves!

Edit: When you provide an open platform for people to collaborate, it can be amazing who shows up and what gets done!

Our World Data website is awesome. It is very generous of Hannah, Esteban, Jaiden and Max to make it available and I can only imagine their passion. I represent a specialist area Mental Illness/Psychiatry and have shared this reservoir of knowledge across the globe amongst peers. Importantly it has aided us in our creation of the CAPE Vulnerability Index which is now getting traction across the globe in particular amongst Foreign Aid doners looking for greater accountability for spend to outcomes. Thank you Twitter @geopsychiatry

I thought that this was what https://www.gapminder.org/ was for, ever since 2005? And yet you don’t mention them?

The work of Gapminder, and Hans Rosling in particular, was a major motivation for the work we do.

In fact, Max worked a lot with Hans Rosling in the past: in this BBC documentary on ending global poverty, for example (https://www.gapminder.org/videos/dont-panic-end-poverty/). You can find him listed as collaborator in the sources (https://www.gapminder.org/news/data-sources-dont-panic-end-p...).

As you say, Gapminder is one organization that takes data from published datasets and communicates it to a larger audience. The Gapminder foundation therefore has a similar mission – promoting fact-based world views, informing people about long-term changes in living standards. But Gapminder is not broadly focussed on global change – it does not cover violence, war, poverty, education, environmental change, etc. – and instead more narrowly focuses on health and demography. The areas of my own personal work: environment, food, and energy, for example, are not extensively covered there.

In fact, for those topics that lie outside of its focus, they often rely on us as input (you can find us referenced throughout their book 'Factfulness' for example). And similarly for us: for long-run trends on aspects such as child mortality or fertility rate, we reference their datasets (https://ourworldindata.org/grapher/child-mortality).

So our mission is very similar and we continue to collaborate, but our approach slightly different. Gapminder is narrower in focus and tries to capture a much more general audience. We go for more breadth and depth, and target a less wide (but still large) general audience.

Do you collaborate with for-profits that do cover these fields (s.a gro-intelligence)?

It would be interesting to see what adding radon would do to the section on air pollution (arguably, radon is a form of air pollution, albeit not particulate).

Since the current estimate in the USA is about 10,000 deaths per year (https://www.ncbi.nlm.nih.gov/pubmed/2118410), with of course, some dissent (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5590909/), this is large enough to show up on the graph, especially for the developed countries.

Unfortunately, getting good data on radon-related death rates is difficult, as it must be parsed out of lung cancer rates with a number of assumptions applied. I figure that getting good data is in your wheelhouse. If not, I could probably put you in touch with someone at AARST or NRSB who may have better data.

Congrats on the Launch:HN!

I like the idea of shifting to a demand-driven model. Unrelatedly, check out https://datausa.io/ if you haven't. Deloitte/MIT created a similar platform for US statistics. Doesn't seem like it's being maintained but it looks pretty/is easy to navigate.

Best of luck. We need it!

That said, I'd like to suggest a book. "The Influential Mind." Long to short, there's more to influence than objective facts.

https://www.amazon.com/Influential-Mind-Reveals-Change-Other...

She gave a talk at Google, for anyone interested: https://www.youtube.com/watch?v=NyME0Idsq9w

> We would really appreciate any feedback you have on what we can do better. Thank you!

What can we do to help you?

Any really honest feedback on the usability of our website would be the best thing for us.

We have lots of content - spanning everything from population growth to plastic pollution, income inequality to cancer rates. This is of course a key part of our work. But this could make it difficult for people to find what they're looking for.

Do you manage to find it easy to navigate? Is there anything you would recommend we do to make it easier for people to find content?

You might find some of the usability testing services useful also

There's a list of some here:

http://www.usefulusability.com/usertesting-alternatives/

and here

https://www.producthunt.com/alternatives/usertesting-2

They give you videos with people's out loud thoughts about for example trying to accomplish specific goals with your sites, which should be really useful.

And for a more guerrilla approach you might find this article useful:

https://medium.com/@leemunroe/how-our-product-design-team-co...

Great recommendations. Thanks for the tips. We've found it really useful recently to sit next to someone in person and ask them to navigate. It reveals a lot of unexpected outcomes just by watching them use the site.

But of course, in-person testing can be tricky to scale so an online service is exactly what we're looking for. We'll check them out!

I think the navigation bar is very busy. It shows up over multiple lines, with different sized and shaded buttons. On mobile/tablet, it takes multiple screen lengths. It seems overwhelming and unfocused.

Perhaps because the navigation is overwhelming, you decided to duplicate your dropdown menu for most of the landing page.

I'm not sure why there is such an emphasis on your media coverage and authors at the top of the landing page. Usually that is a tactic to build social proof for customers feel comfortable buying from a new/unknown startup, but as far as I can tell you're not trying to make a sale.

It took me several clicks to get to an article with an interactive graph -- the cool part. More than just the data, you had good writing in catchy, accessible prose.

My suggestions would be

1) to highlight one cool / recent article+graph at the top of the landing page.

2) Simplify your topic organization; you have 16 different high-level topics, with further subtopics. That relegates you to an unwieldy dropdown. I realize you may be using the SDGs, but I'd prefer 5-6 MECE categories like Environment, Health, Economic Development... into which you can organize SDGs. If you had fewer categories you could fit navigation tabs or make use of UI techniques.

3) Overall, decide on what flow you want from the user and design around it. You have a newsletter, donation, even another web domain for SDG tracking. Who are your primary constituents, and what are you trying to help them accomplish? Is it to delight and educate laypersons, maximize small-dollar donations, highlight the academic rigor of the authors' analysis...?

As a reader, I would come back to the site often if it focused on showing me all the coolest data analysis. I might even browse there instead of Hacker News.

Right now, though, it seems closer in my mind to the World Bank's Open Data website, which has an overwhelming amount of great open data, which I would only visit if I were looking to access a raw dataset.

If lay readership is your primary goal, I would suggest looking for design inspiration from media publications like https://fivethirtyeight.com/ or https://www.newyorker.com/ . Even xkcd.com, which has dated design language, achieves the UX of putting the catchy content up front, and providing navigation for people who want more, rather than the other way around.

Thanks so much for this detailed feedback on the website and our design flow. It's really valuable to get an honest critique of how people find using the site.

These questions/issues of navigation, landing page, structure are all things we're grappling with at the moment and have been making small tweaks then reviewing what difference this makes to how visitors explore.

The rationale for highlighting media coverage and academic citations of our work near the top of the landing was to highlight that we're a trusted source for many publications (and actually from media on both the right and left). Credibility and trust are of course important to our work. But it's interesting to hear it wasn't an important factor for you as a visitor.

Thanks again for your thoughts. We're working on all of this at the moment, and will incorporate all of these suggestions into our discussions of how to do this best.

I think the rigor of the data-driven analysis speaks for itself in terms of credibility and trust, and is the unique strength of your initiative to highlight. Your 'about' page is already very strong to demonstrate your organization and affiliations.

Tufte gives his class in SF once a year I think. I cannot recommend it enough.

This is probably a wacky idea, but...

What if there is a quiz before the start of every article (optional, of course) that the reader can take? Lets say I am interested in child mortality rates, but I don't have any knowledge about it. I probably have wrong ideas too, about the topic. I take a 60 second quiz, then read the article - this will show how misinformed (probably) I am, and then inform me of the topic.

Hey guys, great to see OurWorldInData here, huge fan! I've been following Max and the project for years. One thing I would love to see is "Immigration" as a category and consequently more data about the movement of people.

Hi betocmn, I'm happy to hear that! Yes, took quite a few years to piece everything together.

We have a long list of topics that we want to work on https://ourworldindata.org/about/list-of-all-data-entries (And further work on the existing ones.)

It's not just you, a research project on migration is on the very top of what our readers ask us. Currently we try to make the structure of the site clear and the interactive charts intuitive. But once we get back to research your topic will be one of the first ones we work on. (We are just 4 researchers doing this so it might take some months still though)

Thanks for sharing your findings & datasets, I’ll definitely be following your work. I saw your datasets repo on github, I was curious if any of the raw data or messier bulk data was available?

They seemed to be summarized, or the result of a larger project, vs the structural content.

I did see mentions of using public/gov datasets, but was curious if you had any larger work accumulated?

Definitely not complaining, more curious from a research perspective. Thanks again.

Your site reminded me one of my favorite charts and tweets of all time, so I want back to look and it was from max at ourworldindata way back in 2017! It's this one: https://mobile.twitter.com/MaxCRoser/status/8528130327238574...

Thanks for doing this, I love having big easily accessible data to show people how then world really is.

Wow, this is awesome. I have a project where I want to visualise reported news and weather on the world including trends of news topics. It will be at world.ie (nothing there yet)

There should be many more projects like ours so I'm happy to hear that! I'd really like it if we had as much reporting on the slow long-term changes as we have media that focusses on single events in the last 24 hours.

This is great! Nice job balancing data with explanations. Health data is often oversimplified or not understandable by lay people.

For the cancer section, I'm pretty sure you're showing net survival rates. You should mention this to avoid people from thinking they're actual survival rates.

In the last two charts on the cancer page, what does the circle size show? Total GDP? Patient count?

It shows the population of the country. I agree it should be labelled.

Thank you both for flagging this. I agree it's not clear at the moment. We are doing some tweaks on the interactive charts and will make sure to include this labelling.

This look really interesting. I am really happy to see something with great benefit to society in YC. I would suggest looking at Future Agenda [1]. I have worked with them in the university setting and both OWID and FA seem to be in line.

[1] https://www.futureagenda.org/

This is absolutely fantastic ! I have been banging on about wanting a "ordered backlog" of global issues that could be used to anchor our politics in the daily twitter storms of our attention battles. This looks like it could become my go to site

Good luck guys, we are rooting for you

Thank you very much!! Let us know if you are missing something to make it your go to site.

It would be cool to have the data in a virtual cacheable JS database that can run SQL queries. Eg something that you could just include in a script tag. And also run locally. To allow anyone to easily create and share cool interactive graphs.

Anecdata. Applied for a UN World Food Program grant this morning and they asked us to re-waffle the humans-vs-earth problem. A URL to your data would have been a neat answer :)

this is great, bookmarked it. a news feed of research about the topics i care about it would be awesome. so nice to see progress instead of depressing news articles.

We were thinking of building "a news feed of research about the topics [you] care about".

What do you have in mind? Do you have an example?

We'd be curious what it is that you would expect there. Thanks!

How about an API?

I'm curious about an API also.

One of the most important metrics is entirely missing: wealth and income, finances are totally missing

I think the work you do is vastly important, and it is especially heartening that it originated in academia. In the age of choose-your-own-facts, academics of all stripes have a moral duty to engage the public on what is - and, crucially, what isn’t - known about nature and society. And they’re largely asleep at the switch. So this is major progress.

A couple of long-term ideas:

- breadth of topic: global development trends are a natural and important place to start. However, this is pretty remote, non-experiential stuff for most people. What about a canonical, accessible repository of what is known about nutrition and health, for instance? (“Is bacon really bad for me?” ... with links to every study ever, statistical meta-studies on top of these “research atoms”, and a Vox-esque explainer/lit review layered on top, giving laypeople the tldr). This example is well outside your (well-chosen) scope, probably, but IMO the generic problem here is access to intelligible (scientific) knowledge about anything that affects people and their understanding of the world around them.

- moving the data to the “point of sale”: a huge fraction of the people asking “is the world getting more violent” won’t ever find your website, because people don’t think in terms of “global development trends data websites”, but they just ask questions, and expect Google to answer them. It seems like the long term play is to make expert knowledge itself computational and linked, and for domain-specific knowledge aggregators and synthesizers such as yourselves to work with the Googles of the world so that when I ask “hey Siri, is the world getting more violent?” I get a crisp “no. There are fewer wars and crime trends are declining in most countries, tap to find out more,” with the tap taking me to a source I can trust (because it references said underlying knowledge base) instead of some random shallow news article. I’m in in no way knocking your current product strategy of putting up a high quality space on the web you can curate and control - it’s practical, valuable, and honorable - but I think the future ultimately lies in making expert knowledge structured, and linking it directly to answers to low-intent, everyday questions that everybody has, instead of hoping people know about the domain / are motivated enough to find and dig through your website.

Absolutely agree. The danger of non-profitism is to build something useful and beautiful then forget to tell the world about it. I'm quite shocked I have never heard about OWID, even though I read the "cancer death decline" story only a few days ago.

Yes, you're absolutely right! And actually, much more recently we have been been trying to steer our work towards addressing point (2) you raise above.

It became apparent to us, as you say, that most people will find this stuff simply by asking google specific questions e.g. "how many people are there in the world?". And a lot of our existing content will not appear in Google or Siri as the answer, despite us having it in longer articles on the site.

So we've been very recently trying to answer much more explicit questions that people would ask (although we should have been doing this much earlier), and are thinking of reformatting a lot of our existing content in this way too.

Thanks so much for sharing your thoughts - it's really helpful perspective for us.

Fake news.

I applaud what you're trying to do, and look forward to using it, but the problem is that the most studiously honest broker can be slandered as either the plaything or the cynical agent of vested interests at virtually no cost. The arguments of bad faith actors are not based on substantive disagreement but on inducement of emotional posture.

One strategic response to this might be to formalize your citation process. You cite various media organs that use OWID as a source; could you establish partnerships or public agreements such that if a media outlet uses your material as a source, they do not cherry-pick but must present it in an embed to the reader, and can be called out, cut off, or be subject to contractual penalties if they knowingly misrepresent the source data?

https://ourworldindata.org/about/coverage#coverage

Edit to add that this isn't a problem I think you can or should have to unilaterally solve, but rather than observation that our lack of any good mechanism for certifying sufficient consensus is an unfortunate obstacle to OWID's very worthy mission.

I stumbled upon your website a couple of days ago and found its content extremely useful. The language is precise, opinions balanced and use of data is sensible. It feels like you have no agenda other than bringing some objectivity to the discussions. I hope your journey into the startup world won't change it. Good luck!

Our main goal is what you say. I am happy to hear that. Thank you! It is exactly what we try to do.

Launch HN: OurWorldInData (YC W19 Nonprofit) – Data on World’s Largest Problems