This was a great example of design hindering interpretation. The graph’s diagonal/diamond shape made legibility much harder. A traditionally oriented graph would have been trivial to read and correlation more obvious.

I think they were trying to show something meaningful with the orientation by having their "Trend Line" be the one horizontal metric.

I didn't really understand how they came up with that as a trend line.

I agree that they made an obvious decision to give design preference over legibility. My first glance at the page was "this is neat", but as I tried to wrap my brain around the data, my thoughts changed to "what an unnecessary mess".

I was physically turning my head 45 degrees to the right to figure out what was going on in the graph.

Completely agree. This graph is impossible to read.

You're right, this chart [1] from their blog post about it [2] is far easier to read. I don't like the design as given.

[1] http://2oqz471sa19h3vbwa53m33yj.wpengine.netdna-cdn.com/wp-c...

[2] http://www.visualcapitalist.com/relationship-money-happiness...

I think it was an attempt to prove the point that money != happiness. But, ironically, the trend line shows the opposite: for majority of countries $ and happiness are correlated, hence the trend line.

Agree. But even the data is questionable. How can Singapore & Luxembourg be richer than Norway? Did they include all the Norwegian oil profits?

Yeah - this is nothing short of terrible.

Well, in the attempt to make a snazzy graph the designer has actually obscured what the data show. People are used to the most desirable point being at a corner on this kind of thing, not in the middle of the side.

Regarding what it's showing I have some real reservations about the satisfaction score. It's pretty subjective how happy a person is, and it's not actually that easy to translate the question into foreign languages. Just consider in English:

- Are you happy?

- Are you generally happy?

- Are you satisfied with life?

- Are you overjoyed?

All these things mean something similar but cannot easily be translated to all other languages in the study.

Plus there's a question of how people really think about the question. If I'm living in Scandinavia I will constantly be bombed with the results of this exact type of study. Maybe I won't think so hard when I consider what the answer should be. Because I know it, I read it in the paper.

On the income side, you could argue that the score should be the median income, not the average. There's a number of small, rich countries where a few people are insanely rich. I think Equatorial Guinea showed up on one of those GDP/C lists last time I looked.

I happened to study this in my History degree back in 1995 (I took 'quantification in history' because I liked math, then utterly sucked at it. But I digress).

Happiness is an extremely tricky subject, and the Latin America thing brought back memories of discussions about whether LA countries and cultures simply self-identified as happy, or whether they really were happier. I always thought just asking people if they were happy was a really dumb way of comparing levels of happiness, and that more objective measures of populations, like number of suicides (however problematic in itself, eg a reluctance in some cultures to identify death as suicide) was more reliable.

What this chart shows to me is that GDP is not a good measure of people's wealth. For example Luxembourg is shown as an outlier because it has a very high GDP but not a similarly high satisfaction score. And I think that is because the economy of Luxembourg is largely banking, where the money will just belong to a rich few, not to the average people.

Yeah, it'd be interesting to see how happiness correlates to distribution of GDP rather than size of GDP.

The main problem with Luxembourg’s GDP per capita statistics are cross border workers, who are about 20%. Average and median income are high, but not as much as these figures suggest.

Would like to see this measured with either gdp per capita ppp, or better median wages, preferably adjusted for local purchasing power.

Gdp per capita doesn't take into account different purchasing powers locally. Nor does it take into account income inequality. i.e., the influence of a few people on the happiness measure is negligible compared to their influence on the income statistic.

Singapore having $80k as the baseline is laughable. The median is $30k.

https://en.wikipedia.org/wiki/Median_income

It drops from 3rd to 18th, Qatar drops from second to 23rd.

Yeah, low overall satisfaction isn't that surprising in a 'rich' country if the reality is a crazy-rich aristocracy while the majority live like serfs.

A potentially interesting additional axis would be income inequality. One very wealthy person among the poor can skew the average GDP considerably, but happiness is relative to a fixed maximum and doesn't distribute the same way over the population. Qatar might be an example of this.

How did they measure happiness? If they just asked people, I think this only shows the different answer-cultures in each country.

Considering Brazil and Mexico make up cities that cover the top ten in murder but appear to show positively in this graph, I'm skeptical.

https://en.m.wikipedia.org/wiki/List_of_cities_by_murder_rat...

I've spent significant time in South America and I can categorically say that people are happier and friendlier that the northern americans. Being in a high-crime area doesn't mean you don't have life satisfaction, with your extended family living nearby and your basic needs met, and the fine weather and food.

It's a well known fact that Latin American countries rate themselves as subjectively happier.

Probably http://worldhappiness.report

This is one of the worst data visualization graphs I've seen in my entire life. I had to rotate my monitor 45 degrees to have any idea what was going on.

Iran is on the graph twice..

The graph looks good. Any idea, what did they use to construct it?