Visualizations can provide insight and convey meaning for data. It may be useful to create visualizations to spot trends, outliers, and make the data easier to interpret.

Step 1: What is the purpose of the visualization?

  • Do we need to compare one item to one or more other items?
    • Find the Highest/Lowest value points
    • Show change over time
  • Do we need to show the breakdown of th data?
    • Does the breakdown change over time?
  • Do we need to show how the data is distributed over a certain range?
    • Emphasize groups within data
    • Emphasizing outliers in the data
  • Do we need to show a relationship between two or more variables?
    • Positive or negative relationship?
    • Strong or Weak Correlation?

## Step 2: Determine the types of Data -Numerical data: Numbers that are either discrete or continuous - Discrete: Integers - Ex: Counting units sold - Continuous: Real Numbers - Ex: Temperatures, Revenues

  • Categorical Data: Ordinal or nominal values
    • Ordinal Data: Data with an understood ordering of values
      • Ex: Days of the week, alphabet
    • Nominal Data: No understood ordering of values
      • Ex: Countries, genders

## Step 3: Choose the chart type

  • Comparison
    • Compare items against each other or over time
    • Bar plots, Scatter Plots
  • Composition
    • How parts make up a whole
    • How composition changes over time
    • Pie Chart, Compositional Bar Plots
  • Distribution
    • How the data samples are spread out over one or two variables
    • For Clusters and Outliers
  • Relationship
    • Correlation between two or more variables
    • See trends in data
    • Scatter Plot with Trendlines

Comparison Charts

Column Charts

image

  • Good at showing numerical values for categorical data
  • For example, a table with categories of crops and their respective heights

    Bar Charts

  • Turn a column chart on its side image

We can show production of crops over a series of years: image

The columns are numerical rather than categorical! If we were showing a lot of years or years that aren’t spaced apart evenly, use a line chart: imageimage

Rules

  • Start charts at zero
  • If the labels of the categories are long, use a bar chart

  • Sort items by value rather than name image

  • Time based data should be horizontal, continuous, and run from left to right

Tableau Terminology

  • Dimension vs Measure
    • Dimensional Data: categorical data and ordinal data that are not numerical, axis on charts
    • Measure Data: numerical and suitable for use in calculations
  • Aggregations
    • Some calculation is performed over all records that match a dimension

Create a Comparison Chart

Bar Chart

image image https://public.tableau.com/app/profile/annette.lopez2690/viz/CropProductionBarChart_16318334068440/Sheet1?publish=yes

Grouped Bar Chart

image image

https://public.tableau.com/app/profile/annette.lopez2690/viz/CropProductionSidebySideBarChart/Sheet2?publish=yes