A2-KuangChen

From CS294-10 Visualization Fa08

Let's make a visualization!

Contents 1 Data 2 Question 3 Visualization 4 Discussion

Data

Source: Penn World Table from the Penn Center for International Comparisons of Production, Income and Prices

Formating:

1. Got the data from the data page, saved the text as a csv file
2. Open in Emacs, removed the header info
3. Notice that the columns have no titles
   1. Grabbed the natural text list of columns
   2. Open in text editor, convert to lowercase, delete parens, replace spaces with _
   3. Check for unique column names. Yes.
   4. Regex to add "" around each name; add ,s; unwrap to single row
4. Open in Excel to make sure it looks right
   1. Nope! There were titles, but nonsensical
   2. Remove my own header, and try to figure out what the existing headers mean
   3. They have an appendix for just that. Some interesting fields:
      1. Population: Pop
      2. Real gross domestic product per capita: CGDP
      3. Purchasing power parity: PPP
      4. Standard of Living: STLIV
      5. Group by region: column 7
5. Realized that null values were "na" and Tableau formats as dimension rather than measure. Instead of figuring out how to do it in Tableau, I just search and replaced in the .csv: "na" to ""

Question

Initially: Is there a relationship between the number of people in a country and its standard of life? Next: Turns out Standard of Living is described in the appendix but not available in the dataset. We'll use a proxy: the real GDP per capita.

Visualization

First I plotted some aggregate values all countries over the past 50 years to see a trend in total population growth versus gdp growth versus purchasing power.

Next I want to see how each country fairs over time in terms of "real" GDP per capita. The lines are colored by country. The line size is proportional to the population of the country.

Improving on the previous visualization, I constrained the range in several ways to better use the screen space: set the Y axis to log scale, filter X axis to have the "interesting" years.

Discussion

Tableau is way cool. I wish they had it for OS X. I had to run it on VMWare Fusion and Windows XP. I read the Tableau white paper, but this is the first time I've actually used it. I found the controls fairly intuitive, but requiring a background in the terminology.

I wanted to make a chart with two different X axis, one on the left and another on the right. I did not figure out how to do this in a reasonable amount of time.

---