Life Expectancy and Years of School

Michelle Majstorich

Originally, I decided to just make a scatterplot, as visualized first, to show the data as simply as possible. This I considered to be the least biased that I could make it. Obviously, no matter how I presented the data, it’d be biased by virtue of the fact that I was “presenting,” and that it was presented to me by someone else (the UNDP website). But the scatterplot seemed the purest form, because each data point represented itself (generally), and the axes were labeled. I used the Many Eyes visualizer to do this. I was disappointed by the lack of my ability to customize, meaning the platform I used was biasing my visualization already without my input (other than my choice to use the platform from the start). This scatterplot, though, showed what I wanted it to. I wanted to see the general trend of the data (how the life expectancy increased as schooling did).

My second visualization was my step away from trying to be unbiased. I specifically chose to use region names, such as North America and East Asia (and a few supplied by the data set such as Arab States or Latin America), and averages that either I calculated or that were already calculated for me and were in the data set. I did this because I wasn’t satisfied with some of the region names. Subsequently, the regions are all very general, meaning that whatever the viewer considers to be part of the region will be part of it. After getting the averages, I set up an ageline in Photoshop and typed in region names, changed the font sizes, and eyeballed where they should be on the ageline. I also colored each region name in relation to the yellow North America. The more I considered it to be different, the more opposite I made the color. This was extremely subjective, and I did this to just see if there was some sort of trend, which I suspected there would be (the closer to yellow, the greater life expectancy, and to a lesser degree the years of schooling).

The third visualization was an exercise in misinformation that I stumbled on accidentally. I was determined to create a bubble chart, with a similar slant as my second visualization, but I didn’t want to use Photoshop again, since the assignment was to use more than one platform. I tried Many Eyes and a few other online places, but it didn’t visualize the way I wanted it to. And I realized that my difficulty in visualizing was a problem with the platforms. There was no absolute customization, and the information is misinterpreted. I decided to run with this, and use the misinterpretation. I put my data into Excel and made a bubble chart with as much customization as I could figure out. The result is a chart that makes no sense and gets nothing that I intended across, except the limitations of the lack of customization. The x-axis means nothing, just a numerical value to correspond with region names that the Excel program automatically generated. So, the downward sloping trend shown just happens to exist because I entered the regions in that order.


The fourth visualization was the most manipulated. I took the numbers away and interpreted the data honestly (as opposed to trying to play unbiased by using a chart). This poster has an agenda which I obviously state. Underneath, I state my reason for such an agenda. My simple sentence: “Generally, life expectancy positively correlates with the years of schooling.” is a direct visualization of the data, obviously through my eyes and interpretation. In order to make this assertion, I performed linear regression (basically to see how closely the data correlates). I got an R of 0.74, which shows at least some degree of correlation. This visualization was to show the greatest degree of interpretation, which is also what happens every time someone looks at a chart.

I think that the idea of visualizing data in charts and scatterplots is a way of getting your point across and convincing others of your point without letting them think they were led to their interpretation. That’s why I showed the varying degrees of interpretation. Hopefully, the viewers will realize that each is biased and meant to lead the viewer to a certain conclusion. I think we give too much legitimacy to “data,” myself included, and we should always be aware of this tendency to give so much legitimacy to “data.”

Sources and Online Platforms Used:

“Many Eyes : Scatterplot Showing Trend of School Years an Life Expectancy.” Many Eyes. IBM, 30 Jan. 2012. Web. 02 Feb. 2012. <http://www-958.ibm.com/software/data/cognos/manyeyes/visualizations/scatterplot-showing-trend-of-schoo>.

“Regression Tools – Online Linear Regression.” Xuru’s Website. 2 Feb. 2012. Web. 02 Feb. 2012. <http://www.xuru.org/rt/LR.asp>.

“International Human Development Indicators – UNDP.” Indices & Data | Human Development Reports (HDR) | United Nations Development Programme (UNDP). Web. 02 Feb. 2012. <http://hdrstats.undp.org/en/tables/>.

Advertisement

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s