Principles of Data Visualization

Descriptive Visualization

visualization view that describes a dataset

form of data driven visual communication

utilized as an enhancement or replacement for text based communication

traditionally uses charts, diagrams, and plots

Who invented descriptive visualization?

William Playfair - 19th century Scottish engineer

Who was Florence Nightengale and why is she important?

Chief nurse during the Crimean War

Demonstrated that data could win arguments that sentiment could not

primary objective was to improve troop strength

invented the COXCOMB DIAGRAM to represent mortality data of soldiers.

Coxcomb diagram

looks like a fried circle (slide 4) also known as the polar area diagram

Polar Area Diagram

Technical name of the coxcomb diagram.

uses the 360 degree shape

utilizes the area to represent differences in data categories.

Importing data: how to do it?

connect to a file like an excel file, text file, or JSON file.

connect to a server uses an API to connect to databases and data warehouses like MySQL or Oracle

connect to saved data sources created in Tableau.

Does Tableau have extensive data server connectivity?

Yes

Data Source Extract: How to do it?

Drag datasheets and tables into the connection canvas

Metadata types are automatically assigned for each column.

How to edit data columns?

Field names should be edited for optimal communication

Metadata types should be reviewed and edited.

How to add a datasource filter

Values for a source data column can be filtered before the analysis worksheet.

What is the data pane? what is it used for?

Each data source receives a tab in the data pane.

All of the columns defined in the data source are reflected in the data pane (IMPORTANT!!!!!)

each tab is organized into dimensions and measures

dimensions group or segment the data set

measures aggregate the data set

Dimensions

QUALITATIVE values (customer, product, dates, geographic data)

used to categorize, segment, and reveal details in the data.

dimensions affect the level of detail in the view.

Measures

QUANTITATIVE values that can be aggregated (sales, quantity, profit)

Aggregation is applied to measures by default.

How to convert a measure into a dimension

In some analyses we prefer to have a field function as a dimension even though it is numeric.

Quantity is used as a category of sales types.

Data element properties

Data elements should be renamed to optimize communication

Aliases can be assigned to the values in each field.

Measure number format

Number type can be defined for currency, scientific, or whole numbers can be defined in default properties.

How to assign a geographic role

geographic roles need to be assigned to fields used in mapping

geographic roles include county, state, city, and zip code

What are relationships?

relationships describe how two data tables relate to each other.

based on common fields but the tables are not merged.

orders table contains the field region used as a relationship (example slide 17)

what are data joins?

Data joins merge the two tables into one larger table.

a join is created by selecting open from the drop down in the first table and then dragging the second table.

select the join type and join fields in each table.

what is a data blend?

establishes a common field relationship in a view

establishes a primary data source with a blue check, secondary data source with an orange check.

orange links beside data elements signify the linked field.

Data blend results

data blends combine the totals of two separate queries.

effective technique when source data exists at different levels

What is a data union?

tables for union must have the same structure

select the new union button

drag and drop tables to the union dialog box.

Choosing a visual

create buttons are found in the selection bar at the bottom of the workbook

Analysis worksheets are the base views

Dashboards are constructed from analysis worksheets

stories are constructed from worksheets and dashboards.

How to build a view

data elements can be dragged to the row, column, or center of the canvas.

data elements can also be dragged to the shelves or marks.

How to build a view with show me

command select multiple data elements and then click on the show me tab to see good choices of visualizations.

Adding an axis

additional measures can be added to an axis

the second measure of profit is added to sales in the y-axis

the pill in rows contains measure values representing the 2 measures.

Dual axis

dual axis of Y1 and Y2 can be added by dragging the second measure of profit to the right side of the canvas.

a split pill in rows will represent the two measures.

Visual encoding fields

fields are dragged from the data pane and attached to a type of mark to encode that specific data element

after encoding, marks can be edited in each encoded pill.

defining the marks type

automatic selects the best mark types for your view

selecting unique marks types will tun on additional encoding

shape will turn on the shape visual variable

polygon will turn on the path visual for hurricane tracking

density will turn on heatmap encoding.

encoding points with color

encode the category dimension with the color mark (lil colored dots)

Encoding points with size

encode the sales measure with the size mark (varying sizes of dots)

Encoding Bars with Size

Encode the Sales measure with the Size mark (bars change both height and size)

Encoding with a Label

Encode the Sales measure with the Label mark

A label is supplied for each bar

Encoding Detail

Encoding Category and Sub-Category with the Detail mark provides more granularity to the points

Encoding a Tooltip

Data attributes encoded to a Tooltip display in a list with a hover over the display of a visual glyph

Encoding with Shapes

Encode the Segment dimension with the Label mark

Line Encoding with Path

Selecting the Line visualization style and encoding the Path mark allows creation of a flow map

What does aggregation do?

values for sales and profit can be aggregated into a single point

Aggregating measures

each measure is assigned a default aggregation sum

aggregations for each measure can be changed

aggregating dimensions

use the context menu on sales to create a quick table calculation

select compute using the dimension column like State.

Discrete categories

category based comparison data is presented most effectively with HORIZONTAL BARS, PLOTTED POINTS, and BULLET CHARTS

Product A, Product B, and Product C for example

Discrete categories are BLUE in tableau

Horizontal bar chart

horizontal bar charts work very well for comparison and ranking

encodes quantitative values as length on the same baseline for easy comparison.

Scatter plot

scatter plot encoding provides the correlation between two measures

Bullet chart

combines a bar chart with reference lines comparing actual and target values.

Ordered categories

ordered categories like AGE, INCOME, FARE CLASS, or LEVEL OF SEVERITY

multiples can be presented as a table of charts

ordered categories would also be BLUE in Tableau

for example for a graph, slide 45, bar table with graphs for A and B seperated.

Histogram

histograms displays a distribution separating the data with bins

number of patients are counted in each segment

Ordering a category

select the pill for the category you want to sort like order priority

select sort

select manual

order the categories in the correct logical manner

Continuous metrics

values increasing on a spectrum from left to right

continuous metrics are green in Tableau.

examples include bar charts with time, area graphs, and line charts.

Stacked bar chart

this vertical bar chart stacks sales from three product categories

area chart

defines the quantity of each ship mode by month

Line chart

trend lines are most frequently used to track trend data over time

here we see trends over time by industry sector

How to create a group

group can be created selecting values in the view and clicking the paperclip

new group is added to the data elements

how to create an OTHER group

an other group can be created by selecting the checkbox in the new group data element

How to keep or exclude data points

after selecting data point, right click to keep only or exclude.

add a filter to a view

drag dimension or measure fields to the filter shelf

adding a date filter

drag a date field to the filter shelf and set the time conditions

Adding a context filter

adding a top 10 filter for sub-category

add to context defines technology only for the top 10

creating a set

sets are based on the values in a dimension

all customers with profit over $1750 for example

Set Filtering

sets function as a complex filter on the values of a dimension (not a measure)

Creating a hierarchy

dragging sub-category on top of category creates a hierarchy

Using a hierarchy

clicking on the + and - in dimensions drills up or down

Sort from the toolbar

sort icons on the toolbar will sort ascending or descending

sorting can be engaged on each pill

Sorting

Sorting can be performed using the Y axis of category

Sorting can be performed on the header of east

sorting can be performed on the field label of appliances

Computed sort. what is it?

Tableau offers several computer generated sort options such as ALPHABETIC, BY A FIELD, NESTED, and DATA SOURCE ORDER

Manual sort

a value can be selected and dragged to the desired position in a list

Format views

formatting can be performed for the sheet, rows, columns

the font, style, size, and color for the pane text, header text, tooltip, and annotations can all be formatted to preference

Format marks for presentation

right click on the shape mark with more shapes provides the selection of custom shapes and icons

How to add a trend line

a trend line can be added based on a specified field or right clicking on the view

you can add a trend line using linear, logarithmic, and other model types.

Creating a calculated field

drop down in the data pane provides custom field creation using a formula

calculated fields also serve the role of custom API

Tableau Calculation

Right clicking on a view selects the table calculation

creates a calculated field that computes on the local data

Create a parameter

parameters are dynamic values that can replace constant values in calculations, filters, and reference lines

current date parameter creates a date field that can be using in filters.

Saving a workbook to a PDF

all of the visualizations contained in a workbook can be saved into an Adobe PDF using File → Print to PDF

Exporting an image

worksheets can be saved as an image file using worksheet → export → image.

Export worksheet to data

refined data elements in the worksheet can be exported to a csv file using worksheet → export → data

Lecture Summary

Tableau has a robust selection of files and interfaces for acquisition of data

relationships, joins, unions, and data blends provide several methods of combining data sources.

encoding is the process of binding data elements to marks

visualization worksheets create views of data

visualization views can be formatted, sorted, filtered, and tailored for communication.