I have always had an appreciation for data and its power to inform decisions and drive positive outcomes.
The first career path that I pursued out of college was public accounting at a âBig Fourâ firm. As an auditor I was given specific training in collecting, gathering, and analyzing data to evaluate the effectiveness of internal accounting and financial planning and analysis functions. In most of my auditing experiences I found the client staff to be competent with their analysis and conclusions, yet they often failed to persuade executive leadership to make business decisions that were clearly supported by the underlying data.
As I started to reflect on why these challenges were so pervasive, I observed that the presentation, layout and user interface of the data left a lot to be desired.Â I concluded that if I could improve my data presentation and communication skills it would dramatically increase my value as a professional.
Edward Tufte â My Visualization Role Model
A friend recommended that if I was interested in learning more about the power of well-presented data I should read works of Edward Tufte, a renowned American data visualization expert and professor of computer science and statistics at Yale University. Â Tufte and his analysis of a notorious cholera outbreak changed my life and forever placed me on the path of âpracticing visualization.â
In his short work Visual and Statistical Thinking: Displays of Evidence for Making Decisions, Tufte dissects how Dr. John Snow solved a major health epidemic that took the lives of over 600 London residents in 1854. Dr. Snow used unique data aggregation methods and a heat-mapping layout to illustrate the correlation of cholera outbreaks to well pumps in the city. Through his analysis he discovered exactly which well pump caused the epidemic, and most importantly, prevented further spread of the infection by having the Parish Council remove the handle of that pump.
Tufte then compared this data visualization example against the NASA Space Shuttle Challenger disaster that occurred in January of 1986. The data scientists were aware of the risks that Challenger faced and recommended delaying the mission based on their quantitative analysis and historical stress tests of O-rings. Unfortunately, they were ineffective in the way they presented and communicated the information and as a result were not able to prevent the catastrophe that took place. Tufte goes on to provide examples and recommendations for how the data could have been laid out to more clearly communicate the risks associated with a launch in cold temperatures.
Based on Tufteâs phenomenal juxtaposition of real life data successes and failures, I applied my own professional experience to come up with four best practices to improve data visualization and communication skills:
1)Â Â Â Â Identify the âcustomerâ â Who is consuming the data? Is it your boss, a head executive with deep financial experience, a relatively new person to the workforce with limited subject matter knowledge, or a decision maker that has already made up their mind about what they are going to do? Each of one these âcustomer profilesâ could potentially dictate a different visualization strategy. Know your audience and conform your data presentation strategy to meet their needs.
2)Â Â Â Â Keep it Simple Stupid (KISS) â Although it is not the friendliest acronym, it is one that finance and operation professionals should commit to memory. Simplifying the layout allows the user to better focus on the core message of the data. Â In contrast, complex presentation is difficult on the eye and forces the user to draw conclusions on what data is the most relevant.
For any financial model or analysis I prepare, I always include an executive summary and keep the summary to one page. An executive should be able to make informed business decisions from data. Â An analyst should be able to aggregate their findings and communicate their conclusions in a succinct fashion.
3)Â Â Â Â Be aware of visual design â Visual design should be as important to the finance function as it is to the marketing department. Â In an effort to hone the message and highlight the data findings, the presentation and layout of the document is critical.Â Be thoughtful about color, grids, chart types, and graphs. All of these elements impact the ultimate display of the information and will either better focus the user or will make the message more convoluted.
4)Â Â Â Â Seek feedback from representative customer profile â I highly recommend peer or âlowest common dominatorâ data reviews prior to release.Â If a person with limited background that is unfamiliar with the data can quickly come up with the message/conclusion from review of the analysis, then you know you are on the right track. Conversely, if the reviewer gets confused, overwhelmed or has the wrong takeaway, you will likely have to rethink your work.
Data visualization is important for all business functions.Â Thankfully, we are blessed with great tech companies in our local ecosystem that offer fantastic visualization products and services. Member companies like Tableau Software, Simply Measured, and Moz are leaders in this space and provide great tools for technology and business professionals.Â Implementing best practices and investing in visualization resources can go a long way in taking your data analysis from good to great!