August 2019 – Data Visualization in DH

Blog Post #1

Being able to read and work with data is becoming increasingly important in today’s society. Data literacy is essential as more and more data are collected, used and misused. As more data is collected and used, it can result in more bias and could be misused. As a computer science major, I understand how biased many datasets are and how these biases can influence the world. A biased data leads to biased algorithms and biased people, causing more bias in the world. As expressed in their article, Catherine D’Ignazio and Lauren Klein, most data does not equally represent everyone thus it is inherently biased. This is where data literacy is important as that skill can be used to account for the biases in the data.

When used correctly, data visualization can provide a quick and simple method of reading a complex dataset. However, when misused, data visualization can misinform or even harm society. An example of a harmful data visualization is when politicians misrepresent their influence. The image below misrepresents data by visualizing an increase in petrol price as a decrease.

Tweet from India’s Prime Minister showing the “decrease” in petrol price.

The image below shows another example of a bad or misleading data visualization. The data is represented by a pie chart, where the whole does not add up to 100%. The pie chart leaves out the remaining 35% of the people.

Pie chart attempting to show how much of Skimm’rs have received treatment.

The biases in data and the misrepresentation of data in a visualization are two of the many reasons why data literacy is an important skill in the modern society.

Blog Post 1

Data literacy is important in all aspects of our world. It presents to people statistics, world issues, navigation, and so much more. It provides us with data that can be applied to real- world problems, whether it be in health, inequality, the economy, or the environment. Data visualization renders information in a way that helps communicate the data to the people. However, data ethics are of utmost importance. Bias, accountability and transparency are common issues in data ethics. Withholding bias in certain areas of research can lead to accountability and transparency for all races, genders, and social groups. This can lead to the design of data sets and systems that work towards equality and fairness. Du Bois was an important figure in the beginning of date literacy and visualization. He led his students to collect and analyze data on black communities so that they may think sociologically and go on to compile charts and graphs based on their findings. These charts and graphs could then be used to analyze experiences of black people, especially in Philly, around the beginning of the twentieth century. Their findings showed that black population and fertility rates were increasing.

The following examples from viz.wtf portray misuses of data visualization.

**In this image, the creator violates correct representation of numbers, as the images aren’t directly proportional to the quantities represented.**

In this image, the labeling is not very clear, as the lines are all the same color and there is no clear way to follow them. There are no clear events on the x axis to explain the extreme changes in the y axis either, making it hard to interpret the visualization well.

Practice