Tableau Notes
SUCCESS LOOKS LIKE:
Having a completed “Tableau Desktop Specialist” on my Linkedin Profile
RESOURCES:
Quizzes (with 2012 Superstore data)
INSPIRATION!:
The Data School – VERY COOL GRAPHICS!
BEFORE EXAM:
- Review these notes
- Review screenshots in Class Notes deck
- Review TABLEAU google doc, specifically “What is the exam like”
- Retake Quiz 1, 2, 3, & 4
POSTS:
POSTS WITH TABLEAU TAGS
- daily log 03-08-21 March 8, 2021
- daily log 02-09-21 February 9, 2021
NOTES:
02-16-21
- CSV? Use Text File
- Where is “My Tableau Repository? Inside Documents
- How do I add my own custom shapes? Inside My Tableau Repository, under Shapes. I simply replace the png file with my own.
- What is a role? What’s a type? A type is the datatype. Tableau specifies specific string types, too. Like, global location (e.g. United States is not simply a string)
- USUALLY, tableau will assign numeric data to MEASURES
- USUALLY, numeric data will be in the MEASURES shelf
- USUALLY, qualitative data will be in the DIMENSIONS shelf
-
You can CONVERT numeric to “dimensions” and vice versa
- You can aggregate your MEASURES (your numeric data) – you cannot aggregate your DIMENSIONS (this is intuitive, we cannot find the average of Male and Female. We can find the average HEIGHTS (numeric) etc)
- DIMENSIONS give you discrete headers
- MEASURES is an axis
- DIMENSIONS affect the level of detail (e.g. adding SUB CATEGORY to CATEGORY and getting more of a pivot table view)
- GREEN = CONTINUOUS
- BLUE = DISCRETE
- Numeric data is usually aggregated and the aggregation is NOT ALWAYS SUM – it varies based on your type of view
- For exam, mostly working with DISCRETE DIMENSIONS & CONTINUOUS MEASURES (this again is intuitive, usually there are only two sexes – male or female. Discrete. The weights of these males and females is continuous.)
- example of a continuous discrete is OrderID
- Be VERY CAREFUL to use and double check that you’re using the correct aggregation – SUM for sum and AVERAGE for average
-
AutoGenerated by Tableau:
- MEASURE NAMES: headers in “pivot table view”
- MEASURE VALUES: numbers in “pivot table view”
- Lat + Lon (if geographical data is present)
- QUESTION: how many furnitures
02-17-21
- MAX NUMBER OF TABLES TO JOIN: 32
- In a JOIN, number of columns increase
- In a UNION, number of rows increase
- UNION, needs exact same column names
- JOIN, different Column names
- BLEND Is a join across different data sources
- in a blend BLUE is primary data source, ORANGE is secondary data source
- BLENDS use LEFT joins
- “Field Name” is name in Tableau (does NOT change original data source)
- “Remote Field Name” is the name in the original data source
- see ORIGINAL field names by right clicking on column and selecting DESCRIBE
- You can create ALIASES only from DISCRETE DIMENSIONS (not measures)
ON EXTRACTS
Tableau Data Extract (TDE) is a snapshot of the data that Tableau stores locally. Good for very large datasets of which we only need few fields. Performance is optimised because it queries its own database engine instead of the local data source. When you create an extract of your data, you can reduce the total amount of data by using filters and configuring other limits. After you create an extract, you can refresh it with data from the original data. When refreshing the data, you have the option to either do a full refresh, which replaces all of the contents in the extract, or you can do an
ON WILDCARD SEARCH
You can use Wildcard Search to set up search criteria to automatically include tables in your union. Reference: https://help.tableau.com/current/pro/desktop/en-us/union.htm
ON UNION NAME MISMATCHING
Interesting! You don’t need to know the understanding as that’s covered in the Certified Associate Exam, but still good to remember this. Reference: https://help.tableau.com/current/pro/desktop/en-us/union.htm
SECTION 5: Exploring and Analyzing Data
BAR CHARTS – Useful for:
- comparing data across categories
NOTES:
- How can I change column colors? Drag [whatever I want colored] to the color mark
- How can I create a normal bar chart to a stacked bar chart?
- METHOD 1: Drag [whatever i want stacked] on to the color attribute of the mark panel (Drag a dimension to the color on the marks card)
- “Sales over time?”
LINE CHARTS – Useful for:
- seeing trends over time
- forecasting future values
NOTES:
- default chart is line chart
SCATTER PLOTS – Useful for:
- visualizing relationships between NUMERICAL variables (one MEASURE on columns shelf and one MEASURE on rows shelf)
- Result is analogous to Cartesian chart with x + y coordinates
NOTES:
- Use scatterplots to visiualize relationships between numeric variables
DUAL AXIS CHARTS – Useful for:
- comparing multiple measures
- Analyzing two measure with different scales
- TIP: Quickly create by dragging measure onto the RIGHT side of the graph
COMBINED AXIS CHARTS – Useful for:
- (Same as above)
- Resolves potential misunderstanding issues with the Dual Axis Charts
- TIP: QUICKLY CREATE by dragging measure onto the axis
FILLED MAP
- Size means nothing here BUT Color does!
CROSS TABS / PIVOT TABLES / TEXT TABLES
- Order of pills on the shelf matters here
- Like DIY drag-and-drop pivot tables
- NOTE! When sorting, make sure the correct column is selected
GROUPS:
- create a group to combine related members in a field
- E.G. I have some customers who I’d like to GROUP together
- HOW TO: Right click –DIMENSION– and group (command click who I want )
- Paperclip icon represents group
- TO RENAME: Right click
- FIND MEMBERS helps us quickly select and make groups
SETS – “Like Groups on Steroids!”
- We can do similar things to what we’d to in SQL!!!
HIERARCHIES:
- Order matters
- Save time and prevent “unknowns” using hierarchies
FILTERING:
- We filter by dragging either our measure (continuous) or our dimension (discrete) into the “filter” drawer
SORTING:
- We can use MANUAL sort to change the order of how we view things.
- If it’s a text based field, it defaults to alphabetical
- We can sort by clicking on the “columns” but fastest way is by the “ascending” and “descending” buttons on top
HISTOGRAMS:
- a chart that displays a distribution. It takes a continuous measure and turns it into buckets.
PARAMETERS with HISTOGRAMS:
- What if I don’t want to keep manually changing the bin numbers each time? I can create a PARAMETER (dropdown by DIMENSIONS)
CREATING CALCULATED FIELDS:
- Also, dropdown by DIMENSIONS, we can create CALCULATED FIELDS – eg. Profit Ratio.
- NOTE: it is NOT simply Profit/Sales – we have to FIRST SUM(Profit)/SUM(Sales)
CALCULATED FIELDS FOR DATES:
- ex: ship delay
TREND LINES
- Analytics pane
- To use TREND LINES – both axis MUST be interpretable as a number (Dates are OK – can be interpreted as numeric by tableau)
REFERENCE LINES AND BANDS
- BAND is a shaded portion
SHARING INSIGHTS:
ANIMATIONS:
- format » Animations (off by default)
- Introduced in v 2020.1
DASHBOARD:
- THREE OPTIONS:
- Fixed Size
- Range (scale between)
- Automatic
TILED VS. FLOATING:
- How do we we make something tiled vs floating?
- Tiled by default
- LAYOUT » Floating
HOW TO EXPORT:
- WORKSHEET » EXPORT
- ALSO right click and copy
STORIES:
- “Alternative” to powerpoint
- Each sheet of a story is called a “story point”
- What’s a story? A sequence of dashboards or sheets that work together to tell a story
- STORIES – remembers the actions you do (e.g. selecting)