green happiness

Have you ever wondered what factors influence the overall happiness of the population in your country?

What the world is looking at right now.

Our natural environment, how to protect it, and in particular, how to deal with the causes and consequences of climate change are clearly amongst the leading issues of our times. In late 2018, the well-known activist Greta Thunberg started a school strike demanding the Swedish government reduce carbon emissions. This strike soon went global. In March 2019, around 1.4 million young people across 128 countries took the streets under its Fridays for Future banner, to demand climate action change from their governments [1].

Climate change and environmental protection were the dominant themes of the campaigns, and major international organizations such as the World Bank, ramped up their financial commitments to the environment and natural resource management. Other organizations followed their lead, and started assessing their operation's impact on the environment and incorporating environmental protection into their corporate social responsibility.

How the environment affects people's well-being has also been the subject of academic research. A handful of datasets including indicators of subjective well-being have become publicly available, and can now be merged with other indicators of environmental factors. Most of the studies exploit these datasets to display how people's feelings and life evaluations depend on these factors in their surroundings. These include, for instance: air pollution, land cover, safe drinking water or temperature and precipitation.

In what follows, we first study the happiness indicators around the world and then zoom on the region of Western Europe, showing how different social and economical measures contribute to it. Afterwards, we explore the Environmental Performance Indicators in the same regions and investigate the correlation it has with the aforementioned happiness indicator.

This article explores the major role that the natural environment and conditions play in different country's population happiness.

OVERVIEW OF HAPPINESS AROUND THE GLOBE

The year 2022 marks the 10th anniversary of the World Happiness Report, which uses global survey data to report how people evaluate their own lives in more than 150 countries worldwide. Using their data, we first give an overview of the happiness scores* around the world.

*Note = The score is measured according to the "Cantrill Scale", used by Gallup on their surveys. It's one of the many types of well-being assessment. In this case, it goes from 0-10, where the top of the ladder represents the best possible life for a person and the bottom of the ladder represents the worst possible life.

Overall, we can see the northern hemisphere of the globe scores higher in terms of happiness. On the contrary, Africa, India, as well as the Arab Emirates and Eastern Europe display the lowest scores. This difference is due to the fact that the countries that score higher in the Happiness Report have a higher GDP per capita, social support in times of need, absence of corruption in government, healthy life expectancy, and freedom to make life choices. On the other hand, the poorest countries cannot enjoy these benefits and this has a serious effect on their morale. [2]

Note: By hitting the play button, it is possible for the reader to analyse the evolution of happiness across the years. Hovering over the different countries shows the exact score for each country.

Now, if we select the top 10 countries we can see that within the northern hemisphere, specifically, the Northern Europe countries score the highest, with an average of around 7.5 points.

But why is that? Aren't the northern countries known for being cold, dark, and having higher depression rates?

The most extended factors affecting the happiness of the population come from the social measures taken by the governments. The higher the implication on the citizens the higher they rate their satisfaction in terms of well-being.

Our small study is conducted right in the second happiest country in the world, Denmark. Therefore, we will explore the differences between our home country, Spain, and the one that we willingly decided to live in ;).

Many questions arise when it comes to comparing the happiness of countries. Are you satisfied with your life? Turns out to be one of the very first inquiries that pop up in our heads. To give an answer, using data from The Eurobar surveys [3] we can compare Spain's and Denmark's happiness scores over the years. The following graph shows the share of people who claim to be 'very happy' or 'rather happy'. Spain has constantly maintained its position below Denmark, especially in the last decade, coinciding with the Spanish 2008 crisis.

It is clear that the percentage of happy people in both cases is, on average, quite high. However, to what extent are they satisfied? Stop and reflect for a moment. On a scale of 1 to 10 (with being 10 the most satisfied and 0 being less satisfied). How satisfied are you with your own life? The Gallup World poll [4] asked this question to people in every country.

Following the previous findings, the grade of satisfaction in Spain drastically dropped from 7.5 to 6 when the 2008 economic crisis happened and has failed to reach the same level ever since, wiggling around a grade of 6/10 for the last ten years. Denmark's population satisfaction has also dropped starting from 2008, nevertheless, it has stayed way more consistent across the last ten years, never falling below 7.5.

Influence of social and economical measures on population's happiness

While the previous charts give us an idea of how happiness is distributed around the globe and zooms in between two European countries, we would like to dig further. In this section, we will analyze how different social and economical factors contribute to and correlate with the happiness index.

For a start, we will take a look at the Gross Domestic Product or GDP, which determines the standard of living in a country. In technical words, it is the monetary value of all finished goods and services made within a country during a specific period.

Countries with higher GDPs generate more goods and services, and their citizens enjoy a higher level of living. Thus, a correlation could be established between variables. Does the domestic product concern people enough for it to affect their well-being? By looking at the bubble chart we can observe a positive correlation between both. It is also displayed how in fact Denmark, with a logged GDP per capita of 10.9, scores higher than Spain in terms of happiness. Using the slider on the bottom of the chart we can see how both countries stay consistent through time, and if the reader were to explore further, the size of the bubbles shows information regarding the perception of corruption of the countries. Hovering over the different bubbles displays the mentioned data. Play around!

Focusing now on social measures, the question is analog to the previous one concerning GDP. Is there a positive correlation between the perceived freedom to make life choices and happiness? It may be obvious to the reader that this is in fact, true, and the chart below supports this statement.

Our focus is put on Spain and Denmark, and both follow the trend line for the European countries -the higher the freedom, the happier the people-. However, sometimes, this is not always the case. We believe it is worth mentioning how countries like Portugal (located also in the Iberian peninsula, southern Europe) while having a higher perception of freedom in comparison to Spain (almost 20% more) are in fact around 1 point below in terms of happiness score.

Overall, both economical and social measures positively contribute to the happiness perception among the European population.

Environmental Performance index (EPI)

Let's move on to our second main concern. Sustainability, global warming, climate emergency, greenhouse gases, renewable energy, climate change..., we are sure these concepts sound familiar to you: they have been on TV, in newspapers, in tweets, and in Facebook posts for years now. Climate change is a real phenomenon, and most nations -to a greater or lesser extent - are taking part in the transition to a more sustainable way of life by developing policies and enacting legislation that will contribute to a greener planet.

The Environmental Performance Index (EPI) [5] provides a summary of the state of sustainability around the world by assessing the following parameters for each country: air quality, sanitation and drinking water, heavy metals, waste management, biodiversity and habitat, ecosystem services, fisheries, climate change, pollution emissions... just to mention a few.

In the following world map, we can visualise how the EPI scores are distributed across time and countries, just as we did in the previous sections.

The distinction, in this case, leans towards the differences between developed and developing countries. While Europe, North America, and Australia score higher, South America, Africa, and Asia get the lowest ratings. The explanation behind this may not only be related to the measurements taken by the government or the countries' industry resources to tackle problems like air pollution or water management but also because of the disposition of the developing countries to serve as the factories for the rest of the world [6]

Plotting the evolution of the EPI score across years for each of the continents, we can see the trend is not promising as of right now.

While in 2016 it reached its peak value on average, excluding Europe, it is drastically dropping as we get closer to the present time for the rest of the world.

What happened in 2016? According to the EPI report of that same year [7] the goal to reach carbon neutrality in 2050 shook up the rankings since the beginning of the EPI report in 2014. In 2016 nations across the world invested in plans and took action regarding these measures to tackle the problem. After that, the road to a perfect environmental performance showed its true colors and became less imperative throughout the next years.

Let's not get too sad here, the world may be falling apart regarding environmental terms, but the hope is not lost. Some countries are actually performing incredibly well and could serve as an example for the rest.

If we take a look at Denmark's performance across the years, we can see its fast evolution rate, ranking 4th in 2016, 3rd in 2018, and reaching number one in the year 2020, while not even appearing on the podium for 2014.

Note: We encourage the reader to play with the "Year button" to check how the Top10 and Bottom 10 countries changed over the years! :)

Spain's performance, on the other hand, is way less impressive.

Although from the year 2014-2016 its score was even higher than Denmark's, it has been dropping ever since, only to acquire an EPI of 75% in the year 2020.

Moreover, an interesting evaluation can be established if we compare the environmental performance with one of the economical measures that we have found contribute highly to happiness, the GDP. Let's look at the plot below:

It can be seen that, once again, an increase in the GDP per capita correlates positively with the EPI. There are multiple assumptions that can stem from this conclusion. Higher economical power defines how much money a country can allocate for the implementation of measures that will reduce its carbon footprint and elevate its EPI. Thus, Denmark's high GDP and strong involvement in sustainability matters reflects in this chart. Spain is not far behind, although to a much lower extent.

The poorest performers have problems that extend beyond their inability to sustain environmental and human health.

These nations show that environmental performance is an issue of governance – only well-functioning governments are able to manage the environment for the benefit of all.

Happiness vs epi -the final evaluation-

Finally, after studying how the social, economical, and environmental measures separately contribute to happiness, we want to see the overall correlation of the environmental performance index to the happiness score 😁.

If you have followed the different plots across this article, you have probably noticed how every aspect has had a positive correlation to the happiness score. It would therefore be weird if the EPI behaved differently. But luckily it doesn't!

Focusing on the year 2020 we can see how the higher the score in the EPI, the happier the population. With a star marker, you can see our two favourite countries, Denmark on top, and Spain on the bottom.

So yes, indeed, environmental performance does help to increase the overall happiness of the population.

A hesitant reader could be thinking, very angrily, that correlation doesn't mean causation. And we totally agree. Our whole study is based on the assumption that having an environmental-friendly country increases happiness and looks for support for the hypothesis. But keeping in mind that there are a lot of factors contributing to this equation, we just wanted to open a path that could be explored further, and the results are shown here - even though promising- are not conclusive nor hold the absolute truth.

Don't go just yet!

Can we determine the happiness score based on the environmental variables?

Oh no... a machine learning model again? Yes. We are AI students and we love to investigate our problems by feeding data to a computer and seeing if it can outperform a human.

In this case, we wanted to see if the computer could guess the happiness score for a country (0 to 10) based on the EPI indices. This can help us see which variables a machine determines as most important to the prediction, and therefore, give another idea of what to look at or improve to achieve higher or lower happiness terms.

For this, we trained a Random Forest Regressor model [uh, oh.. an ugly technical term] and got a Mean Squared Error of 0.30 (this is the average distance between the prediction and the true value)

Is this a good score? Well, we can plot one by one the predictions and the true values for our test data and visualise the behaviour of our model:

Okay, we see that the predictions follow in some way the values for the true values, but let's do one more thing: let's try to do classification.

For this, we discretized the continuous output variable "happiness index" turning the values of [0-10] into 5 separate classes as follows: "very unsatisfied", "unsatisfied", "mildly satisfied", "satisfied", "very satisfied" . We then created a Random Forest classification model to try and predict those values based on the previous EPI indices. The results are shown in the following confusion matrix [9]:

What does this mean?

The accuracy for this classification is 0.52 which, for a 5-class labels classification, is not that bad, however, is neither good to make a robust classification.
As seen in the confusion matrix on the left, the model performed well when predicting the very unsatisfied class (6/9 right guesses) or the mildly satisfied class (8/11) but bad for others more abstract classes, like very satisfied.
The lack of samples with the class very satisfied might have been a factor for which the model did not classify well enough for this label.

Feature importances

Lastly, on the side is shown the feature importance for the models, the two more important features are H2O and EPI.

The H2O index summarises how well a country provides healthy drinking water and sanitation. The EPI takes all the other indices into account so, as expected, this feature is one of the most important.

A priori, by looking at this plots, we could say that healthy drinking water is one of the most important features contributing to the happiness from an environmental perspective. Nevertheless, a more extensive benchmarking should be done to reach a more stable result.

Note: Just 5 features have been selected to display in the image, the rest of the 6 features ended up having around 0.05 importance.

In the end, both models -regression and classification- have a very average performance. This is probably because the data on which they have been trained is not numerous and/or unbalanced between classes, for instance, there is a sample for each of the 138 countries, and after splitting for train and test, the training data resulted in just 92 observations.

Note: There are some resources to cope with this lack of observation such as regularization but are out of the scope of this analysis.

In conclusion, a model could actually help us predict the happiness score... but for now, let's just rely on our human judgement and societal context for each of the countries.

overview of both the epi and the happiness score worldwide

Have you noticed we love maps? At this point hopefully yes. It would be sad to go without a plot that summarises the happiness scores and the EPI indices around the globe, so here you have it [10].

Note: To start, click on the layer button in the top right corner to change the map view, and explore around ;)

Not convinced with our conclusions? Would like to explore even further?

Don't hesitate to download and play with the source code!

Click here to see the code! 👍🏽

References

Meet the authors

Lucía

Madrid, Spain

s220492

Javier

Madrid, Spain

s182105

ANna

Barcelona, Spain

s202436

Page updated

Google Sites

Report abuse