Guide To Data Visualization – Part 1

Data visualization is a key component in the data science and analytics process that not only guides you in building a better model but could also lead you to build a report or a dashboard that an end-user can consume on a daily basis. Data visualization is an art by itself as not every chart or a table on a plain white screen can be called a good chart. In this post and the following post, we will see what are some of the do’s and don’ts while building a dashboard or a data viz.

Most of the key points were taken from Stephen few’s books – Show me the numbers & Information dashboard design. We will first begin with table design. Any dashboard usually either consists of a chart or a table, and even a table has appropriate ways to be displayed so that consumers grasp the information rightly. In this post, we will go through some of the key aspects of a proper table design that shows the quality of a well-designed dashboard.

Table Design

A table is a structure for organizing and displaying information and a table exhibits the following characteristics.

  • Information is arranged in columns and rows.
  • Information is encoded as text (including words and numbers).

It is the arrangement of the information that characterizes tables, not the presence of lines that visibly delineate the structure of the underlying grid. Tables are not exclusively used to display quantitative info. Whenever there is more than one set of values, and a relationship exists between values in the separate sets, you may use a table to align the related values by placing them in the same row or column. For example Meeting agendas like the below;

TimeTopicFacilitator
9 AMOpening RemarksScott
10 AMDemoDoug
11 AMDiscussionDavid
12 PMPlanningKatie
Fig 1

When deciding whether to use a table or graph to communicate your quantitative message, always ask yourself how the information will be used: 

  • Will you or others use it to look up one or more values? Then it’s a prime candidate for expression in a table.
  • Will you use it to examine a set of values to discern patterns or trends? Then, use a Graph.

When to use tables?

Here are some key points to consider before using a table in your dashboard or report;

  • The display will be used to look up individual values, but not the series or comparison (categorical).
  • Precise values are required. 
  • Values have more than one unit of measure (like gallons and dollars in the same view).
  • Both Summary (Avg, Sum, count etc.) and details are needed to be displayed together.

If your data has one or more of the above, it’s better to go for a table.

Tables are commonly used in data visualization to display one of five potential relationships, which can be divided into two major types: 

1. Quantitative-to-categorical: 

A Quantitative variable is a variable that displays a numerical value (integer, decimals, floats etc..) and a categorical variable is a variable that has many values under the variable/column and is commonly a non-numerical value.

  • Between One Set of Quantitative Values and One Set of Categorical Items – Below is an example of displaying a quantitative value against a categorical value.
SalespersonSales
Rob$1500
Mandy$1000
Terry$2000
John$1500
Total$6000
Fig 2
  • Between One Set of Quantitative Values and the Intersection of Multiple Categories.
SalespersonJanFebMar
Rob$100$200$150
Jones$150$200$100
Total$250$400$250
Fig 3
  • Between One Set of Quantitative Values and the Intersection of Hierarchical Categories
Product LineProduct FamilyProductSales
HardwarePrinterPPS$ 1000
PXT$ 2000
RouterRRZ$500
RTS$250
SoftwareBusinessACT$ 2000
SPR$1000
Total$6570
Fig 4

2. Quantitative-to-Quantitative:

  • Among One Set of Quantitative Values associated with Multiple Categorical items.

The image might look similar to the one above, but the comparison made is vertical, (ie.. comparing values of a categorical variable for the same month in this case.)

SalespersonJanFebMar
Robert100015002000
John250010001500
Total350025003500
Fig 5
  • Among Distinct Set of Quantitative Values associated with a Single Categorical Item.
SalespersonSalesReturnsNet Sales
Rob15002001300
Mark25005002000
John200010001000
Total600017004300
Fig 6

So, what we saw above are some of the ways a table can be displayed and are the ideal ones to use in a dashboard or a data visualization report. Before we move on to the overall design practice for a table, let’s see what are the variations in a table design.

With the general structure of columns and rows, tables can vary somewhat in design. Structural variations can be grouped into two fundamental types:

  • Unidirectional – Categorical items are laid out in one direction only (Either across columns or down rows).
  • Bidirectional – Categorical items are laid out in both directions.  

Overall table design practice:

  • Delineating columns & rows:
    1. Use white space alone whenever space allows.
    2. When you can’t use white space, use subtle fill colours.
    3. When you can’t use fill colour, use subtle rules.
    4. Avoid grids altogether.
  • Arranging data:
    • Columns or Rows:
      1. Arrange a set of categorical subdivisions across separate columns if they are few, and the maximum number of characters in those subdivisions is not too large.
      2. Arrange time series subdivisions horizontally across separate columns.
      3. Arrange ranked subdivisions vertically down the rows.
    • Groups and breaks:
      1. Use just enough vertical white space between groups to make groups noticeable.
      2. Repeat column headers at the beginning of each new group.
    • Column Sequence:
      1. Place sets of categorical subdivisions that are arranged down the rows of a single column to the left of the quantitative values associated with them.
      2. Hierarchical relationships must be placed from left to right.
      3. If categorical subdivisions have a meaningful order, sort them in that order.
  • Formatting text:
    1. Avoid text orientation other than left to right(horizontal).
    2. Keep numbers to the right, keeping the decimal points aligned as well.
    3. Keep text on the left.
    4. Align dates however but maintain consistency.
    5. Place a comma to the left of every 3 whole-number digits. (example 7,125,000,999)
    6. When negative numbers are enclosed in parenthesis, keep the negative numbers themselves right-aligned with the positive numbers.
    7. Express days as 2 digits.
    8. Express years consistently as either 2 or 4 digits.
    9. Express months either as 2 digit number or a 3-character word.
    10. Use a font that is legible and use the same font throughout.
    11. Boldface, italicize or change the colour of fonts when useful to group or highlight.
  • Summarizing values:
    1. Make columns containing group summaries visually distinct from detail columns.
    2. Consider placing summaries in the group header if the information extends down multiple pages.
    3. Repeat column and current row headers at the top of each page.

These are some of the important points to be followed while building a dashboard or a data visualization report which could play an impactful role in how the end-users consume it. A good visualization speaks a lot and with BI playing a pivotal role in every organization, it’s vital to follow proper visualization standards to create an impact. In the next post, we will look at how to design graphs. Until then, Happy reading!

Share the Gyan!

Related Posts

This Post Has 3 Comments

Leave a Reply

Your email address will not be published.