Basic Data Concepts
Fields
Qualitative Fields (Dimensions)
- Describes or categorizes data
- Tells you what, when, or who
- Slices the quantitative data
Quantitative Fields (Measures)
- Numerical Data
- Provides the measurement for qualitative category
- Can be used in calculations
Row-Level Record
Understanding how to answer the question "What does a row of data contain?" empowers you to ask questions about the data you encounter.
Some datasets can have more granular data encapsulated in each row than others. As data becomes less granular, it is referred to as aggregated data. The granularity or aggregation in a row afects the questions we can ask of the data, and the discoveries we can make.
Aggregation, Granularity
Granularity in Tablaeu is increased by adding dimensions into the view.
Pill Types
Formating of Pills
- Dimensions come out to the view as themselves, Measures come out to the view as aggregates
- Discrete pills are blue, and continuous pills are green
Views
- Bringing a continuous field into the view creates an axis
- Bringing a discrete field into the view creates panes with labels
- Bringing a continuous field into the "Color" style creates a gradient, while a discrete field provides a separate color for each item
Maps
<type of field> | Dimensions | Measure |
---|---|---|
Continuous field | Symbol map | Filled map |
Discrete field | Color Palette | Color Gradient |
Dates
- Can be brought into the view as continuous or discrete
- If the date icon in the pane is blue, it will be treated as discrete
- If we right-click drag, we can choose how to view the date (year, month, etc.)
- Continuous date truncations are treated as an ongoing progression along an axis
- Discrete date parts are treated like categories, where each category member has its own graph and trend line
Reading Common Chart Types
Visual analytics leverages our pre‑attentive attributes, which are the visual cues humans automatically process with sensory memory. We can notice and interpret these kinds of attributes quickly and without special effort.
Overview of Charts
Elements of Charts
- Filter
- In Tableau, you can click on a filter to view a specific subset of data
- Mark
- Mark is the term used to describe the visual representation of the data. In a bar chart, the mark is called a bar
- Tooltip
- Tooltips can be found when you click a mark and will provide more information about the data the mark represents
Questions to ask when looking at charts
- What does this chart represent?
- Does this chart show any particular patterns or trends?
- Is this all of the data?
- Is it clear what has been measured, and what the numbers represent?
Misleading Visualizations
- Bar charts with axis not including zero
- An axis without zero can be misleading on a bar chart, as it can distort the scale of differences between categorical data
- Color confusion
- Color can effectively draw attention to or differentiate different areas of data, but can cause confusion if not used carefully. e.g. instead of visualizing density as lighter to darker, doing the opposite (lighter means denser, not usually how we interpret color)
- Wrong chart type
- Not all chart types work for the selected dimensions and measures
Chart Types and their Uses
Line
View trends in data over time.
Example: Stock price changes, website page views during a period of time
Bar
Compares data across categories.
Example: Volume of shirts in different sizes, percentage of spending by department
Heat map
Show the relationship between two factors
Example: Segment analysis of target market, or sales leads by individual rep.
Highlight table
Show detailed information on heat maps
Example: The percentage of a market for different segments, or sales numbers in a region
Treemap
Show hierarchical data as a proportion of a whole
Example: Storage usage across computer machines, comapring fiscal budgets between years
Gantt
Show duration over time
Example: Project timeline, duration of a machine's use, availability of players on a team
Bullet
Evaluate performance of a metric against a goal
Example: Sales quota assessment, performance spectrum (great/good/poor)
Scatterplot
Investigate relationships between quantitative values
Example: Male versus female likelihood of having lung cancer at different ages, or technology early adopters' and laggards' purchase patterns of smartphones
Histogram
Understand the distribution of your data
Example: Number of customers by company size, student performance on an exam, frequency of a product defect
Symbol maps
Use for totals rather than rates. Be careful, as small differences will be hard to see
Example: Number of customers in different geographies
Area maps
Use for rates rather than totals. Use sensible base geography
Example: Rates of internet-usage in certain geographies
Box-and-Whisker
Show the distribution of a set of the data
Example: Understanding your data at a glance, seeing how data is skewed towards one end, identifying outliers in your data
Questions that Lead to Viewing Correlations
When deciding whether you need a scatter plot, consider the following questions:
- Will the user be interested in whether a variable increases or decreases with the changes in another variable?
- Will the user be interested in determining whether a change to one variable has an effect on another variable and what that effect is?
- Will the user need to know how closely a variable follows the changes to another variable?