The human brain is an exceptional detector of visual patterns. This is one of the fundamental reasons why it took so long to develop a computer that could beat humans at chess; it was difficult to get computers to recognize patterns as proficiently as someone like Garry Kasparov.
The trend in data analytics is to try to leverage this characteristic. The idea is that the easier it is for people to see a picture of their data, the easier it will be for them to detect important patterns. Seeing trends in a graph is a much more intuitive process than inferring a trend based on statistics. This is why company dashboards have become so popular. They exist to get as many eyes as possible to look for patterns in the company’s performance data.
The human brain also has a propensity to turn patterns into stories and narratives. This is useful because narrative stories transmit the meaning of complex information efficiently. The business world is very excited about combining visualizations and narratives to communicate the actionable implication of complex data.
The objective of this course is to teach the usage of Tableau to maximize the efficiency and efficacy of data analytics projects. The objective is not merely to teach how to use Tableau. In keeping with this objective, the course will be organized according to the life cycle of a data analysis project instead of according to the features of Tableau.
Professor Jana Borg is at pains to point out that while the human brain is very good at seeing patterns, its also very good at making up patterns and seeing what it wants to see instead of what’s actually in front of it. This is “confirmation bias.” In order to cope with this tendency, it is best to enter the visualization part of an analysis project with a fixed plan for how the analysis project will unfold.
So, an ancillary goal of this course is to teach a strategy for organizing analysis projects. First, the plan is distilled into carefully considered questions and hypotheses. Then, the data is analyzed and visualized. Finally, it is distilled into a narrative story that will be readily understood by other people. This process is called a structured analysis plan, and will be elaborated in coming sections. An overview of the steps involved is outlined below.
- Craft the right questions
- Design and implement structured analysis plans
- Create important graphs in Tableau
- Transform data in Tableau and publish dashboards
- Tell data stories
- Design effective slide presentations
The Science of Data Analysis
The goal of data analysts should be to arrange the insights of their data in such a way that everybody who seems them is able to clearly understand their implications. John Stuart Mill, quoted below, suggests that that is the realm of science.
Science groups and arranges its truths so as to enable us to take in at one view as much as possible of the general order of the universe.” John Stuart Mill
One of the main jobs data analysts have in today’s world is to automate business processes and insights using predictive computer algorithms. Don Knuth, quoted below, states that is the aim of science too.
Science is knowledge which we understand so well that we can teach it to a compute… Don Knuth
Definitions like these are why the title “data scientist” has become so prominent in today’s world of analytics.
With this said, there is also an art to data science. That art is figuring out how to arrange the things we learn about our data into something that has meaning. In other words, the art is in crafting a narrative, or telling a story, that fits the scientific truths.
Practically, the way to begin this process is to look at the data in many different ways until you figure out the order the data’s insights should be organized into. This is what visualizing data does. It helps determine what the data means, in a very efficient manner. Properly visualized, data helps you zero in on the facts that should be taken into consideration in your statistical models.
Tableau is powerful data visualization software. Whereas other tools, like Excel, might require minutes create visualizations of large datasets, Tableau can visualize the same data in seconds. Tableau is also unique in that it automatically uses best practices in visualization science to format the graph view; these capabilities come from the fact that its founders were world leaders in visualization science from Stanford.
Tableau is big-data ready. Whereas Excel crashes beyond 1 million rows, Tableau can handle huge volumes of data. Under the hood, Tableau runs SQL queries. Given these big data capabilities, Tableau has built-in capability to connect with most kinds of databases directly.
One of the unique features for which Tableau is famous are how easily it can be used to create “dashboards.” Business intelligence dashboards give you views of key business metrics and indicators in real time; their goal is to allow anybody looking at them to immediately know whether something is going wrong or right with the business. Dashboards are frequently designed by an analyst, but then published to people all over the company so everybody can keep an eye on them.
There are a few different versions of Tableau, each of which has different capabilities. There’s a free version called Tableau Public, which allows you to do most things the other versions allow, but which limits your queries to 1,000,000 rows. The only limitations are a requirement to save your data to the Tableau website and only connect to Excel spreadsheets, text files, and Microsoft access files. There are also personal and professional paid versions of Tableau desktop. These allow you to save data and workbooks locally. The professional version adds the ability to connect to databases. Finally, the professional version also can connect to Tableau server, which is a means of sharing things locally at your own company in a secure and controllable way.
This content is taken from my notes on the Coursera course “Data Visualization and Communication with Tableau.” It is part of the “Excel to MySQL: Analytic Techniques for Business Specialization” specialization.