How Happy Does the Sun Make Us?


Final Projects

Introduction

I decided to continue exploring happiness scores for the final project since I saw a lot of opportunity to dive deeper into the data. Originally I had looked at happiness scores for countries around the world along with how factors such as GDP, life expectancy, etc. played into this score. Although this was interesting to examine, this approach has been taken countless times. For the purpose of this project, I decided to explore some of my own questions which came up while working on Lab 4.

I had recently come across the infographic shown below which ranked various cities around the world by the average annual sunshine hours. This inspired me to think about sunshine in relation to happiness.

Figure 1: Infographic presenting average annual sunshine hours around the world

It’s common knowledge that Vitamin D is positively correlated with mood and wellbeing. In fact Seasonal Affective Disorder is a type of depression that occurs annually during times of less sunlight and colder weather. Since sunlight and weather can play such a big part in a person’s happiness, I decided to investigate whether there is a correlation between how much sunshine a country receives and the happiness of the people who live there. I was interested to see if some of the happiest countries were also the ones that received the most sunlight during the year and vice versa. Although I was initially interested in this question, I was open to letting the data guide my exploration.

Method

Datasets

I continued to work with the world happiness scores dataset from Kaggle that I had previously used in Lab 4. This dataset contains happiness score data for 147 countries from 2015 until 2022. The happiness score is a decimal number between 1 and 10 that is computed based on the responses to the Gallup World Survey. This survey asks participants to rank their current life on a scale from 1 (most exceedingly bad conceivable life) to 10 (most excellent conceivable life). Based on the answers to these questions, a national average is computed which is considered to be the happiness score of the country. Along with the numerical score, this dataset also includes information about how a set of six defined factors affect this score. However, for the purpose of this project I was only concerned with each country’s over all happiness score.

Figure 2: Snapshot of happiness score dataset

Since I knew I wanted to explore the correlation of happiness with sunshine, I searched for a dataset that contained information about sunlight duration for countries around the world. I ended up finding a dataset on Kaggle that contained this information. This dataset breaks down sunshine duration per month for various cities in countries around the world.

Figure 3: Snapshot of sunshine dataset

Since the happiness score data is at a country level, I aggregated the sunshine duration values by taking the average of the sunshine duration per month for all the cities within a country. One thing worth mentioning is that this method may not result in the most accurate information since countries like the United States have vastly different sunshine durations in different parts of the country. For example, the variation between the sunlight received in Alaska versus Florida is lost in the aggregation. Regardless, this dataset still serves as a good starting point to explore the correlation between sunlight and happiness.

Since both datasets contain information broken down by country, it was easy to join the two on this field.

Tools

To visualize this data I used Tableau, a software used to create interactive data visualizations. I was able to easily import the data and join the two datasets together in Tableau.

Process

My approach for this project was to create two basic map visualizations to present both datasets. Based on the results, I would then determine how to dive deeper into the data. Although initially I was interested to see whether happiness and sunshine duration were positively correlated, new questions started to come up as the data unfolded. During my exploration I became interested to look beyond just total sunshine duration, but rather break this down further and explore whether the correlation may actually lie in the variability of sunshine over the course of the year. This process of exploration is reflected in the layout of the final visualization. The initial maps are the focal point of the dashboard and are supplemented with a variety of additional charts and graphs.

After the first pass of creating a dashboard to present happiness and sunshine data around the world, I conducted user research to understand the usability and effectiveness of the visualizations. I then incorporated this feedback and created a second and final version of the dashboard.

User Research

Target Audience

Since this project presents global data about happiness and sunshine which are both universal topics, the target audience does not fall into any one specific demographic. However, since I was working with global data I was keen on getting an international perspective during my user research. I was fortunate enough to recruit one participant who resides in the UK and a second participant who resides in India. Beyond just an international perspective, I was able to get feedback from someone from a “happier” country as well as a “less happy” country.

Process

After asking a series of pre-test questions to each participant in order to get a sense of their background and familiarity with the topic, each participant was asked to complete a series of tasks. The tasks were as follows,

  1. Which country has the highest happiness score in 2022?
  2. Which country has the lowest happiness score in 2022?
  3. Which country receives the most sunshine in a year?
  4. Which countries stand out to you the most in the bubble chart?
  5. Do you notice any trends in the sunshine duration hours for the happiest countries versus the least happy countries?

I requested users to think aloud as they completed each task in order to get a sense of their thought process while navigating the dashboard. Upon completion of the tasks, I ended the session with a post test questionnaire to understand what they liked and didn’t like overall.

Findings

I noticed that users were utilizing the hover action a lot in order to find the most happy and least happy countries on the map. Additionally, the shading of the countries on the map was vital since they immediately knew to take a closer look at the darker/lighter countries depending on what they were searching for. I noticed the users making comments about the relationship between the happiness score and sunshine duration for many of the countries. This further solidified the effectiveness of presenting the two maps together on the dashboard. I also received good feedback about the colors chosen to represent the data. Users enjoyed seeing happiness scores associated with the color green since green usually indicates something positive. Similarly, sunshine data associated with the color orange aligns with people’s schemas of sunshine and added to the readability of the dashboard.

Another valuable finding from the user testing was that users seemed to be searching for quicker ways to find the data they wanted to see. For example, when asked to find the happiest country, I heard statements such as “It would be nice to see a list of all the countries based on their happiness, or at least the top X countries and the bottom X countries.” I incorporated this feedback by including a supplemental chart alongside the maps. Users also seemed to want to quickly find the countries from tasks 1 and 2 in the bubble chart. I leveraged the ability to add annotations, shown in Figure 4, in order to clearly mark both Finland and Afghanistan which were the countries with the highest and lowest happiness scores in 2022.

Figure 4: Example of use of annotations based on user’s feedback

Lastly, users seemed to have some difficulty in identifying trends when the line graph displaying monthly sunshine duration information was cluttered with too many lines. Based on this feedback I decided to create two separate line graphs instead of one which is shown below in Figure 5 (use the slider to see the before and after).

Figure 5: A before and after of line graphs comparing sunshine hours in the most happy and least happy countries

Results and Findings

After incorporating all the feedback I received, the final dashboard to explore happiness and sunshine around the world can be seen here as well as below.

As mentioned above, the maps are intended to be the focal point of this dashboard. Since the data is location based, a map was the most effective way to visualize both datasets. Each country is shaded a certain color on a monochromatic color scale to represent the value of the metric being presented. This allows the user to gather a good amount of information in just a glance. The most interesting finding from the two maps is the lack of a positive correlation between happiness score and amount of sunshine. For example, as shown in Figure 6, you can see that the Scandinavian countries which consistently have the highest happiness scores year after year actually don’t receive that much yearly sunshine.

Figure 6: A comparison of happiness and sunshine duration for Scandinavian countries

On the other hand, the African and Asian countries centered around the equator which receive the most amount of yearly sunshine fall on the lower end of the happiness score scale as shown in Figure 7.

Figure 7: A comparison of happiness and sunshine duration for African and South Asian countries

Although the two maps serve as an effective method of exploring the relationship between the two datasets, I wanted to include a single visualization that combined both metrics for each country. A bubble chart, shown in Figure 8, worked well for this situation since I was able to represent a country’s happiness score with the size of the bubble and it’s yearly sunshine duration with color.

Figure 8: Bubble chart representing happiness and sunshine duration per country

Immediately I noticed that the largest bubbles (happiest countries) were mostly located towards the inner part of the bubble chart as shown in Figure 9.

Figure 9: Selection of the “happiest” countries

I began to wonder whether instead of happiness being correlated to total sunshine duration, it was instead correlated to the variance in sunshine received in a year. For example, places like Sweden and Norway have cold and dark winters and receive almost no sunshine during these months. However, on the first warm day of the season when the sun is finally shining, people in these countries experience feelings of pure joy and excitement for the coming season. This contrasts with the inhabitants of countries in which there is less variance in the amount of sunshine year round and arguably no overwhelming feelings of joy at the turn of the season.

To explore this hypothesis I decided to break down the sunshine data by month and plot the duration for each month in the year on a line graph. I specifically focused on comparing the countries with some of the highest happiness scores (Finland, Iceland, Switzerland, Denmark) with the countries with the lowest happiness scores (Botswana, Zimbabwe, Afghanistan). The results are presented below in Figure 10.

Figure 10: Sunshine duration per month

Both graphs are presented using the same x and y axis which allows a fair comparison to be made between the trend of the variance in sunshine duration. It’s interesting to see that the happiest countries all have a similar trajectory of number of hours of sunshine during the year. There is a large increase of sunshine duration in the middle of the year compared to the least happy countries which exhibit smaller changes in sunshine duration during the year.

Although it is hard to say how much of a factor the variance in sunshine duration really is on happiness scores around the world, these visualizations present the possibility of a correlation.

Conclusion

Ultimately, happiness is a complex concept to capture in a single number to represent an entire country. Not only do a variety of factors significantly impact this number, there are also many ways in which these scores may be biased. For example, knowing what we know about the relationship between weather and mood, it would be insightful to know when the survey was given to participants in each country. Respondents in one part of the world may have completed the survey during their winter while others may have completed it during their summer. A participant’s current situation can easily skew responses about their happiness. Another thing to note is that happiness may mean completely different things in different parts of the world which makes it harder to make an exact comparison between countries.

Although the happiness score can be biased and impacted by a multitude of factors, it still garners a large amount of interest each year when the annual World Happiness Report is published. These scores continue to have the ability to provide a peek into the similarities and differences in the human experience across the world.

References

  1. https://www.kaggle.com/datasets/prasertk/sunshine-duration-by-city
  2. https://www.kaggle.com/datasets/mathurinache/world-happiness-report?select=2022.csv
  3. https://worldhappiness.report/
  4. https://medium.com/@joychurin/world-cities-ranked-by-average-annual-sunshine-hours-c7b631aef4bd