Project 3

Gapminder review

Although Gapminder offers mainly classic orthogonal representations, it can present a significant quantity of information in a single chart (different variables: 2 axes, bubble size, bubble color), while in a dynamic mode it adds the ‘time’ variable offering a really picturesque and descriptive representation. It also offers much freedom for customization/exploration.

Dataset exploration

I decided to explore data concerning world population's incomes and financial inequalities. The chart below represents an overview of wold's wealth distribution. The full pie is the total income of the world's population, while the different colors represent the economic classes of the world in five groups of even population (20% of the whole population each). The numbers are averages starting from the first data available in 1979. Only a quick look at this chart renders the wealth distribution inequality obvious. The 20% richest part of the world possesses almost half of planet's total wealth.

But then another interesting question arises. Which are the relations of these 5 economic classes during the dynamic changes of wold's wealth distribution? To examine this question I made a series of motion charts each time with different economic classes at the axes. The first one shows the relation of the two lowest classes over time.



As we can see, the bubbles tend to form a linear pattern with a positive relationship, meaning that as the poorest class increases its income share, the 2nd poorest class does the same. So they should both be drawing their extra share from some other class(es).
So, the following chart shows the relation between the poorest and the middle class.



The linear pattern is not as clear as in the previous chart but it is still apparent with a slightly smaller upward slope. The poorest class continues to increase its income share as the middle class does the same. So they should both be drawing their extra share from some other class(es).
In a similar way, the following chart shows the relation between the poorest and the 2nd richest class.



Here the linear pattern has almost completely disappear. The bubbles seem to be placed almost in a random way, however still in a slight upward slope.
So the last chart shows the relation between the poorest and the richest economic classes.


At this point we can see a clear inversion of the previous trends. The pattern becomes again clearly linear but for the first time the slope is downward. As the poorest class's income share is shrinking, the richest class's rises. Poorest people's income share loss seems to be transformed into excess income share mostly for the richest people.
But what about the relation of the richest class with other classes than the poorest?


As we can see the pattern between the 2nd poorest class and the richest one is the same as the previous, only with a slighter downward slope.
Following are the two remaining couples (middle-richest and & 2nd richest-richest).

The above trends continue in respective manners. The linear pattern continues to exist with a descending slope until it becomes unclear but still downward in the last chart (2nd richest-richest).

In conclusion, the above examined charts reveal that as in societies the richest people increase their wealth share, this becomes in a direction of inequality, meaning that in such a case the lower classes people loose part of their income share in proportion to how poor they are. Or put in a different way the poorer they are, the more they loose.

Project 2

As part of World Values Survey’s cultural context, I chose to visualize a survey question concerning people’s views on wealth. I was curious to find out how important do different societies consider the material wealth. Apart from the nationality of the survey’s respondents, I thought it was interesting to combine it with another variable, such as the educational level of the respondents. Hopefully this could offer a valuable insight into how education affects people’s perspective towards the significance or vanity of the state of having a lot of money and expensive things.

The visualization I chose is a "donut" chart type with multiple charts on the same page representing each country separately. The source code can be found here. The adjustments required for the present project's case were:
  1. Acquiring the data of interest from World Values Survey website and adjusting them to csv format with countries as a new value instead of states and the scale of importance instead of the old age group values.
  2. Separating the above modified data into 9 different sets of data, each for the respective educational level of the respondent and 1 set of the total (all educational levels) survey results for all the countries.
  3. Distributing the new csv files to new folders with respective index files and indicating their context by entitling them.

Below are the links of the visualizations in an ascending order of educational level attained. Last one is the visualization of the total numbers. Wherever the charts are missing is due to lack of available data for the specific variables.
Ideally, all these visualizations could be integrated on a single page, where the user should be able to have a much higher interaction with them. For example the page could be equipped with a slider representing the various educational levels on a constant scale. As the user would adjust the "education" scale, the "donut" charts would be live transformed accordingly. This could offer significantly more interactiveness and usability. Unfortunately, such implementation was beyond my capabilities given the time limitations.

Project 1

According to the students’ self rating of their skills, I made the following visualization for each one of them. I tried to place the skills in a logical array around the polygon with the most relevant skills close to each other, in order to have as far as possible a smooth distribution and render the “skill areas” more apparent to the viewer. I also made sure that the diagram scale has a 0 minimum and a 10 maximum for all cases.

Students' skills

After that, I tried to classify the students into according the field, that each one is stronger. This first rough classification was the following.

Programing: Students 4, 6, 14, 22, 24, 27
Artistic: Students 1, 9, 10, 11, 16, 20
Math/Statistics: Students 3, 5, 21, 25
HCI: Students 12, 15, 19, 23, 17, 26
Neutral: Students 2, 7, 8, 13, 18

By observing the above polygons I tried to match the students into groups of 3 in a way that each one’s skills complete the others’. These are the groupings I made based on that observations.
The goal I wanted to achieve through these visualizations was to form groups that are interdisciplinary, which in my view is a strong advantage of a team as it broadens its capabilities. On the other hand I wanted the group members to share common interests because I think it would reinforce the team’s unity and a friendly atmosphere of collaboration, which is a very important factor in a team’s success as well.

Groups' skills

As we can see above, in most cases the polygons tend to complete each other, as in areas of small cavities of a student's polygon another student's of the same group attempts to cover it with a respective peak.

With regards to my second criterion I transcribed the students' answers into labeled fields of interest. Then I made some bar graphs displaying these interests in a descending order of popularity. The different colors represent the different students.

Students' interests

Unfortunately, I didn’t find a way to combine my two criteria because the skills and the interests seemed in most cases to coincide. This means that one student’s skills are similar to their interests, which makes it difficult to combine the interdisciplinarity with the common interests.