Data & Statistics: Why It Matters
Phone2Action fellow explains — with a literary twist
My name is Yonathan Dawit, but I go by Yon as it easier for others to say and spell. I like Yon because it has plenty of amusing derivatives like Yonboi and Yonbaby, or any other playful iteration that a future friend has yet to create. Now, how was that for an intro? Would, ‘I was born free, but everywhere I am in chains’, work better? That Rousseauvian response may not answer the question, but it sure is pretty! Although bringing up Rousseau may open up a rabbit hole of politics I don’t want to fall into just yet. Who knows what kind of beasts lurk there? Perhaps a Leviathan?
My interest in political theory aside, I work as a data scientist for Phone2Action as part of our Civic Tech Fellows program. This program is an extension of the Civic Tech Fund that Phone2Action established in 2015 during the White House Demo Day. My job at Phone2Action is to look at our production data, client data, customer support data, the relationships between the numbers (numbers can get lonely too), and determine their causes and significance.
Excitement for working in data sounds a tad strange, I know. Most people regard statistics as just a series of numbers, some of which may be somewhat interesting, or unexpected. I mean numbers don’t lie, so there can’t be too much to them right? It’s true that numbers don’t lie, but by golly can they be deceiving, and this is the core of data analysis and statistics.
Sure the work is somewhat technical, but there is a spirit to the method that I would like to share, that anyone can adopt. When looking at a number with an analytical lens, we can’t take it at face value. We have to question it, and understand what it is composed of. I challenge you, my reader of this unorthodox blog post, to question the statistics you see on a daily basis, whether it be on your TV, computer or your phone. Take the data on smart phone acquisition for example. Here is an interesting fact. Fifty percent of adults around the globe own a smartphone. That is quite an astonishing fact, and would seem to indicate some sort of inequality, but it does not tell us a whole lot. This is the challenge. I want us to question this number, assuming the number itself is valid. When was this statistic calculated? How was it done? How are smartphones distributed in the world? Can we assume they are distributed evenly (Please don’t! Assumptions in statistics are dangerous if done incorrectly!)?
What we could do is look deeper and dig up new numbers, such as how for example, South Korea had a smartphone ownership rate of 88 percent in 2015, whilst Kenya had an ownership rate of 26 percent in 2015 (both numbers come from survey data). The natural question that comes to mind is, why do these rates differ so much? A statistician would ask himself — are we comparing apples to apples? Are we considering all the factors? Perhaps not. A possible factor could be access to education, with a suitable metric being literacy rates. There is a statistically significant chance that there are numerous other factors we could come up with regarding the 50 percent rate. What we should take away is that when looking at a statistic, we should not think of it as a number, but rather as a story. The actual number itself is just the title of the story, and the questions we ask about it fill out the pages. What I love about my fellowship is that I get the chance to dig into data of millions of people who have engaged in advocacy efforts. Data seems so removed, but in advocacy, these numbers represent the human ideal of affecting change. So, how are you using data to effectuate change that matters?
