Tableau Steps Flashcards
The steps to do something in Tableau
Fixed Set
In Viz
Open Sheet > (In visualization) Ctl+click & drag over data > hover over data > click ‘linked circles’ icon, click ‘Create Set’
Dynamic Set
In Data Pane
Open Sheet > (in Data Pane) right-click dimension >’Create’>’Set…‘>Conti in General/Condition/Top tab
Filter Set
In Filter Pane
Open Sheet > (in Filter pane) Click dimension > ‘Create Set…‘>Conti in General/Condition/Top tab
Combine Sets
In Data Pane
Right-click set (in Data Pane)>’Create Combined Set’>choose how to combine sets (looks like Full, Inner, LeftOnly, RightOnly joins)
Join
Data Source > drag table to canvas>
double click table/right click table>click Open (or tables if you already have a set) (Note: skipping this step creates a Relationship) >
Drag in other table (next to first table)
Union
Data Source > drag table to canvas>
double click table/right click table>click Open (or tables if you already have a set)(Note: skipping this step creates a Relationship) >
Drag in other table (below to first table)
Blending
Upload 1 Data source
> (in Data Source tab) Drag table from 1 data source into view > upload and drag for 2nd data source
> Input data from both data sources into the sheet (paperclip will automatically appear on a potentially shared dimension/measure)
EDIT Blended Relationship:
Open Worksheet>Data tab>’Edit Blend Relationship’
https://www.youtube.com/watch?v=ZkeRWhj7CCk
Relationship
Data Source > drag table to canvas>
Drag in other table (next to first table)
Create one Show/Hide option for all filters on a Dashboard
Open Dashboard > create horizontal/vertical container > place filters in container > Select container (can do this by clicking the drop arrow of any of the filters and select “Select container: horizontal/vertical container”) > Drop down arrow > “Add Show/Hide Button”
How would you see which gender contributes the most distance, per respective Fitbit category?
Chart should show the percentage of people that were male/female per each category.
Be mindful of using low granularity.
Dimensions:
Distance (Whole Number)
Gender(dimension)
Category(dimension)
1- Create Fixed LOD Calculated Field:
2 - Create another calculated field with a ratio:
3 - Assemble visual:
1- Create Fixed LOD Calculated Field: Distance per Category
{ FIXED [Category] : SUM([Distance]) }
2 - Create another calculated field with the ratio between summed [Distance] & [Distance per Category]: % of Category
SUM( [Distance] ) / SUM( [Distance per Category] )
3 - Assemble visual: Drag Gender & Category to rows and ‘% of Category’ to columns
See pic in Notes
Create a chart that shows how many people continued logging steps after their first day.
Fields:
User ID
Activity Date
Steps (Only Measure)
Cohort Analysis
1- Create a Calculated field,
‘First Day with Steps’ :
IF [Steps] > 0 THEN { FIXED [User ID] : MIN ([Activity Date]) END
2- Create a Calculated Field,
‘Days with steps’:
IF [Steps] > 0 and [Activity Date] > [First Day with Steps] THEN [Activity Date] END
3- Create Calculated Field,
“Days from first steps”:
[Days with steps]-[First Day with Steps]
4- Assemble visual:
Column = [Days from first steps]
Rows = [First day with steps]
Label & Color = CNTD([User ID])
Filter out Nulls (right click > exclude)
Marks = Squares
See Notes for image
You only have ‘Steps per hour’, ‘Activity date’, and ‘User ID’ fields, but we want to show the average steps per day per user.
How do we do this without showing an ‘Activity date’ on the canvas?
Create a new calculation with an INCLUDE LOD called ‘Average daily steps’
{INCLUDE [Activity date]: [SUM([Steps per hour])]}
Add AVG([Average daily steps]) and [User ID] to Canvas (Text and Rows).
See notes for image.