Tableau Best Practices To Improve Performance

Tableau Best Practices To Improve Performance

1. Data source

·         Keep analysis simple, use subset of data or create an extract.

·         If your data is not good for analysis then do resize/ETL process.

·         Use join properly otherwise u will get duplicate records or u may lose data.

·         Always use star schema. If dimensions are extend into more than one dimensions then create view in your database then go for joins. (Example- product _type-> product_sub_type ->product then make it into a single view)

·         Hide/remove unused (Dimensions/measures) in order to increase Query time or minimize extract refresh.

·         Use extract whenever you face performance issue.

·         Avoid using costume SQL if there is a need then only use.

·         Create Indexes on all database column which are part of joins and/or Filters.

·         Avoid ‘NULL’ values at DB level.

·         Data Connection: Go with Extract-> for files, slow performance Dbs.

§  Go with Live-> High performance DBs (Frequently data changes). If live connection is giving slow performance then use tableau server as data source.

2. Calculations

·         Use calculated field carefully. For example - Try to use Boolean calculation

IF [Order Date] = TODAY () then “Today”

ELSE “Not Today” END

Rather than using above calculation use bellow calculation for faster

[Order Date]=Today ()

·         As like above example Use ELSEIF single word rather than using separate ELSE IF (Called Nested IF).

·         Always use Today(), never use Now(), (it compare the value in the time stamp level and degrade the performance).

·         Creating calculation field from multiple data source is not recommended in tableau because Tableau individually query each data source separately to retrieve the values for your calculation so it will time taking process.

·         In calculation try to avoid Table Calculation because table calculation compute at Tableau level

·         If there is a complex calculation then do it in the database level.

·         Instead using ad-hog grouping (default grouping) use calculation.

3. Filters

·         First apply the filters then Build the Visualizations.

·         Always avoid using Exclude option

·         Limit the use of Quick Filters. Use Dashboard ‘Use as Filter’ option or create action manually.

·         Avoid selecting “Only Relevant Values” for your quick filters. This requires sequential queries. Do not use this when not needed.

·         Avoid high-cardinality quick filters (multi-select or drop-down lists). High-cardinality quick filters are slow to load and render.

·         Avoid performing filter on measures.

·         Use context filter properly. Limit the use of Context filter coz it’s create temp file.

·         During publishing workbook don’t publish with multiple Value (list), always recommended compact list (dropdown list).

4. Formatting

·         Use same Text/number/currency format to all reports.

·         Use same/proper Color/Color Legend in all reports.

·         Use Indicator or KPI color as (red, green, and yellow).

·         Tooltip formatting will give your view good looking and understandable.

·         Use Annotations if required.

5. Dashboard

·         Before creating dashboard ask your end-user what type of device they are using, then design the dashboard and fixed the size (i.e. laptop, desktop, mobile, blogs etc.) as per the device they are using.

·         Use Dashboard layout containers and Dashboard object.

·         Limit the number of worksheets on a dashboard, (4 to 5 sheets).

·         Use Actions instead using Quick filter and make your dashboard interactive.

·         While using filter action it recommended to select Exclude All Values.

·         Avoid too much complexity.

·         Avoid too many KPI/Alerts.

·         Avoid inappropriate visuals.

·         Avoid Mismatch between information and its visual representation.

6. Testing

·         Compare tableau reports data with underline data if it is same then your Report is working fine if not then check at tableau level (for example a calculated field is not giving exact data)

·         Start performance recording so you will get SQL code and other things.

·         User’s interactions (i.e. Click, filter, parameters, chart selection, actions etc.).

·         Raw data load and refresh

7. Images/symbols/Logos

·         Use png rather than jpg or other.

8. Others

·         Never place any unwanted field to marks card because they will add more level of details (or they will add more group by clause).

·         Check your filters, parameters, sets, calculated fields are working as expected or not.

·         Chose Graphics/visual as per types of data.

·         Chose related sheets for your dashboard.

·         Remove unnecessary data from tooltip.

·         Avoid creating large cross-tab data.

·         You have geographical dimensions like country, state, city, postal code   and you are not creating any Map the then change the geographical default roles to None because tableau create automatically latitude and longitude. So it will take time to load data and may degrade performance.