Data visualization history

Editor's Notes

• #2: Hello and thank you for being here. I am Amy Germuth, an independent evaluator in education located in Durham NC. This presentation serves as an introduction to data visualization and to this session in particular.
• #3: So what is data visualization? Her are just some examples – some good, some great, some neither.
• #7: More precisely, the Mandelbrot set is the set of values of c in the complex plane for which the orbit of 0 under iteration of the complex quadratic polynomial z_{n+1}=z_n^2+c
• #9: Tarek Azzam and Stephanie Evergreen define data visualization as follows:
• #10: The use of quantitative or qualitative data To produce a representative image Which is readable by viewers That supports exploration, examination, and communication of the data Thinking back on the examples I just showed you we can see where some better reflect this definition than others.
• #11: Let’s talk a little about the history of data visualization. Its history is important since it can help us understand where and why data visualization emerged, where it has been, and where it is headed.
• #12: Broadly speaking, the roots of data visualization include cartography / mapping, statistics, data, visual thinking, and technology, among many others.
• #13: The first known quantitative display is from about 950 – it’s a multiple time-series graph showing the changing position of planets and moons.
• #14: Nicole Oresme was the first to use bar graphs to demonstrate a variable that depends on another valu Although William Playfair is generally regarded the inventor of modern charts
• #15: based on the publications of his books Commercial and Political Atlas and Statistical Breviary which included the first line, bar, and pie charts. Part of what made these books so accessible was the development of the printing press by Johanne Guttenberg in the 1450s.
• #16: Time and again, as we look at the history of data visualization, we will see how technological improvements push data visualization forward.
• #17: Also notable is that it took until the 17th century for someone (Renee Descartes) to visually represent data using 2-dimensions, providing the basis for the Cartesian coordinate system we use today.
• #18: In the 1800s there was an explosion of data displays – Geologist William Smith developed the “Map that Changed the World” which depicted different strata of rock in Great Britain as emanating from a single universal strata, supporting the Darwin’s new theory of evolution. Additional displays addressed social issues such as Dr. John Snow’s 1855 dot map
• #19: It geographically identified the location of cholera deaths in relation to the infected Broad Street pump, supporting the hypothesis that cholera was spreading due to contaminated water. In 1857 Florence Nightingale used circular charts to show that more British soldiers died of poor hygenic conditions in the battle field hospitals than in combat during the Crimean War.
• #21: This is Charles Minard’s graph of Napoleon’s Russian Campaign which shows the decimation over time and geographically of Napoleon’s army during the 1812 campaign. There were few new data visualizations introduced in end of the 1800s and early 1900s due to the public’s lack interest in quantification versus display. However, by the 1960s there was an emergence interest among academics of the visual display of data, led namely by John Tukey and later Edward Tufte.
• #22: Additionally, emerging technologies including personal computers, and specifically the Apple Macintosh, first available in 1984 and which focused on graphics, pushed us forward in the field of data visualization.
• #23: Now we can manipulate and interact with data visualizations, as was dramatically shown by Hans Rosling in his 2007 Ted talk.
• #24: Here he is using bubble charts to depict life expectancy trends over time for multiple countries and cultures.
• #25: Data visualization in evaluation is not new.
• #26: In 1997 Gary Henry edited a New Directions in Evaluation issue titled “Creating Effective Graphs: Solutions for a Variety of Evaluation Data” This year, Data Visualization and Reporting Topical Interest Group founder Stephanie Evergreen published “Presenting Data Effectively”
• #27: I bought a copy this morning. And let me tell you – it’s flying off of Sage’s display. And if you look really closely on page 78 guess whose picture was included? So how has data visualization been used specifically in evaluation? Azzam and Evergreen identify four major ways. They are:
• #28: Understanding Collecting Analyzing Communicating Understanding includes understanding organizations, programs, change, etc.
• #29: Understanding includes understanding organizations, programs, change, etc. Graphic facilitation is just one data visualization method to do so.
• #30: Here’s an example of GIS mapping which can be used in early phases of evaluation to better understand populations, needs, resources, etc.
• #31: Visualizations have allowed us to better understand how to improve the quality of data that are collected. This is an area of increasing interest among survey designers. Including Don Dillman and Jon Krosnick
• #32: Again, understanding is a major reason for the use of data visualization in evaluation. Here’s an example of a word tree being used to better understand connections among qualitative data.
• #33: Data dashboards are also of increasing interest as they allow us to see program performance on multiple indicators at once, potentially showing new relationships that may otherwise be missed. Data visualizations have also been used in communication – including results, patterns, and theories of change. Perhaps the one we are most familiar with is the simple logic model.
• #34: However, data visualizations have limits.
• #35: These include
• #36: Data quality – again, it is the old adage – garbage in –garbage out. The quality of the data displayed is always limited to some degree by the quality of the data included. It is the evaluator’s responsibility to share all limitations of the data he or she is displaying. Causation – data visualization is often used to depict relationships. Again, we must be careful not to imply relationships and / or causation where it does not exist. Supporting? Do data visualizations actually support understanding, collecting, analyzing, and communication? If they do not then one must ask what are their purposes? In what way do they support evaluation? Are they just pretty pictures? To close, Data visualization has a long history – but it will most likely have an even longer tail. There is still much to learn about visualizing data and how data visualization can be used to support evaluation efforts. But that’s for my co-presenters today to discuss.
• #38: Thank you so much.