A2-HannesHesse
From CS294-10 Visualization Fa07
Contents |
[edit] The relationship between wealth and meat consumption
For this assignment, I planned to look at the development of various countries' economic output (measured in gross domestic product) and meat consumption. I acquired data about the meat consumption per capita between 1961 and 2002 as well as data about the development of these countries' populations and gross domestic products in that time from Swivel.
I had imagined plotting small multiples for these time series: Two rows would show the meat consumption per capita and the GDP per capita over time, each row containing graphs for each country. For various reasons (see below), I did not achieve this.
So instead, I restricted myself to figures for three discrete years: 1961, 1981 and 2001.
[edit] Visualization 1
As a first visualization, I displayed bar charts showing the meat consumption and GDP per capita of 2001 next to each other. This provided a first overview, as well as a relative ranking of the different countries' meat consumptions.
[edit] Visualization 2
I then created a scatterplot of the two variables, each point representing a country. I would have preferred the country names nexts to the dots instead of in a legend. The scatterplot gives a quick overview of the ratio between GDP and meat consumption as well as the absolute GDP and meat consumption per capita and therefore fulfills two purposes. The outlier Japan was quickly identified, while a great number of other countries were quite similar.
[edit] Software critique
Using Tableau was a frustrating experience. Several factors made it incredibly hard to use:
1) I had data on world population by country over time and meat consumption by country over time. The rows contain the countries while the columns contain the years from 1960 to 2005. Each cell then contains the country's population or meat consumption. I could find no way of importing this data correctly into Tableau. Tableau wants to treat each column as an attribute or dimension. Thus, it thought my dataset had as many dimensions as it contains years, where in reality it is only three-dimensional.
2) I then decided to strip down my data. I selected the 13 countries with the highest total meat consumption in 1961 and the data for the years 1961, 1981 and 2001. I also included three columns with those countries' populations for these years. This makes for a total of 7 dimensions (country, 3 x meat consumption, 3 x population). In addition, modifying the underlying data was painful because Tableau did not allow a data file to be simultaneously opened in Excel. So to look at the original data, I either had to every time close the Tableau file and open the data file in Excel or create a copy of the data file.
