A2-RagnarSkulason

From CS294-10 Visualization Fa08

Contents 1 Step 1. Pick a domain that you are interested in 2 Step 2. Pose an initial question that you would like to answer 3 Step 3. Find a database that has the data you need to answer your question 4 Step 4. Create the first visualization 5 Step 5. Is there a correlation between HDI and wine consumption? 6 Step 6. Make the visualization readable and show the correlation at the same time 7 Step 7. Same visualization with bar chart 8 Step 8. demographic locations against wine consumption and HDI 9 Conclusion

Step 1. Pick a domain that you are interested in

I think this was the most difficult problem, which domain to visualize. After some browsing and data exploration I decided to make wine my domain, and especially wine drinking throughout the world.

Step 2. Pose an initial question that you would like to answer

My domain is wine and wine drinking around the world, my best initial question is:
Are wine consumption and stages of development correlated?
Or in other words, is wine such a luxury product that only more developed populations drink it?

Step 3. Find a database that has the data you need to answer your question

The databases I used were:
Wine Institute
Human Development Reports (HDI)
One PDF had to be converted to text before I could use it. I found the tool Able2Extract which solved that problem perfectly.
As all the data was in text format, Excel was used to crunch the numbers to fit. No magic was used there.

Step 4. Create the first visualization

First things first, create graph with countries vs HDI (Human Development Index) and wine consumption

There seems to be some correlation here, but it is not clear enough to be certain. Note, the label 2005 on the X-axis and the captions mean wine consumption in 2005. So next question would be:
Is there a correlation between Human Development Index (HDI) and wine consumption by countries?

Step 5. Is there a correlation between HDI and wine consumption?

This time I will use scatter plot and add a correlation line to it.

There we have it, there is a correlation in these. But the graph is quite ugly and not well readable.
So, next task would be to make the visualization readable and show the correlation in the same time.

Step 6. Make the visualization readable and show the correlation at the same time

I switched axes and distinguished the countries with color hue:

Much better, though you cannot map countries with the color hue it works by distinguishing the points. How about using bar chart?

Step 7. Same visualization with bar chart

The same graph but using a bar chart

I think this is what I was looking for.
But how about graphing it by using demographic locations against wine consumption and HDI?

Step 8. demographic locations against wine consumption and HDI

Here wine consumption is graphed with a circle and HDI magnitude is marked with color.
The bigger the circle the more wine consumption and red low HDI while green is high HDI.

This image concludes the data.

Conclusion

My graphs did explain lot of information about wine consumption and I thought I had figured the trend. The more developed the country, the more wine consumption. That is what I did read from my later graphs. But when I did plot wine consumption against geographic location I concluded that there was no correlation in wine consumption and development. The wine consumption is mainly in Europe, and a few countries to which Europeans have spread their wine culture.

---