Free Salesforce Tableau Consultant practice questions

Question 1

At Northern Trail Outfitters, a published data source's Connected Workbooks tab shows 15 workbooks, but its Tableau Catalog lineage lists only 12 workbooks downstream. The data steward asks the Tableau consultant why the counts differ. Which two explanations are valid? Choose 2 answers.
A. Lineage excludes any workbook saved in a Personal Space, so those workbooks are missing from its count
B. The tab counts only workbooks whose sheets use a field from the data source, while lineage counts every connected workbook
C. Lineage counts only workbooks whose sheets use at least one field from the data source, unlike the tab
D. Lineage counts only certified workbooks, while the tab counts certified and uncertified workbooks
E. Lineage counts only assets that Tableau Catalog has indexed, while the tab counts every published asset

Accepted Answer

Correct answer: C. Lineage counts only workbooks whose sheets use at least one field from the data source, unlike the tab | E. Lineage counts only assets that Tableau Catalog has indexed, while the tab counts every published asset
Catalog lineage counts workbooks that have been indexed, connect to the data source and use at least one field in it, with worksheets that use a field, while the Connected Workbooks tab counts any workbook that connects to the data source (whether or not fields are used) that the user can view. Catalog can count only indexed assets, whereas the server counts any published asset. Workbooks in a Personal Space are still counted by Catalog, with sensitive metadata replaced by Permissions required. Lineage counts do not depend on certification, and the claim that the tab requires field use is backwards: the tab counts connected workbooks whether or not any fields are used, while lineage requires at least one used field.

Question 2

Ursa Major Solar's data engineering team must clean, pivot and combine installation and service data using a visual, repeatable process, and then run the resulting flows every night on Tableau Cloud without anyone opening a desktop tool. The company is on Tableau Enterprise. Which combination of Tableau capabilities should the consultant recommend?
A. Build the flows in Tableau Prep Builder and keep them running on a laptop with Tableau Reader installed
B. Build the flows in Tableau Prep Builder and schedule the published flows with Tableau Prep Conductor
C. Build the flows with Data Interpreter in Tableau Desktop and schedule them with Tableau Bridge refreshes
D. Build the flows in Tableau Catalog and schedule them from the lineage page of each published data source

Accepted Answer

Correct answer: B. Build the flows in Tableau Prep Builder and schedule the published flows with Tableau Prep Conductor
Tableau Prep Builder is the tool where flows are built to combine, clean and shape data, and Tableau Prep Conductor schedules and automates published Prep flows on Tableau Cloud or Tableau Server as part of Data Management, which Tableau Enterprise includes. Data Interpreter in Tableau Desktop only cleans files such as Excel sheets, and Tableau Bridge brings data from private networks into Tableau Cloud rather than running flows. Tableau Catalog provides lineage, discovery and data quality warnings, not flow authoring or scheduling. Tableau Reader is a free viewer for Desktop workbooks and cannot run flows.

Question 3

At AW Computing, a security review revoked the right of the Tableau service account to create and drop temporary tables in the Teradata warehouse. Since then, several live dashboards have become noticeably slower. What should the Tableau consultant recommend to the DBA?
A. Restore permission to create and drop temp tables and check that there is sufficient spool space for the queries
B. Grant the service account full database administrator rights so that Tableau can manage the warehouse tuning
C. Move the workbooks to custom SQL connections, because custom SQL avoids any need for temporary tables
D. Keep the restriction and add more data source filters, because temporary tables are only used for extracts

Accepted Answer

Correct answer: A. Restore permission to create and drop temp tables and check that there is sufficient spool space for the queries
Tableau creates temp tables to help improve performance and add functionality, so if creating temp tables is restricted, workbooks and views are not as fast as they could be; Tableau also asks whether the environment has sufficient spool space for the queries. Temp tables support live query performance, not only extracts. Custom SQL usually runs slower because Tableau cannot optimize it. Granting full DBA rights goes far beyond what Tableau needs and is not the documented recommendation.

Question 4

Cumulus Financial's finance team exports a branch scorecard from Excel. The sheet starts with a report title and a run date in the first three rows, and the column headers are split across two rows: a year header spans several quarter sub-headers. The Tableau consultant must decide how to make the file usable as a data source. Which two approaches should the consultant recommend? Choose 2 answers.
A. Load the sheet into Tableau Bridge first, because Bridge removes introductory text from Excel files before refresh
B. Connect to the file in Tableau Desktop, turn on Data Interpreter, and verify that the headers were flattened correctly
C. Connect without changes, because Tableau treats the title and date rows as metadata and ignores them automatically
D. Define a named range in Excel that excludes the title block and connect Tableau to that named range
E. Use the two header rows as is, because Tableau builds a year and quarter hierarchy from stacked column headers

Accepted Answer

Correct answer: B. Connect to the file in Tableau Desktop, turn on Data Interpreter, and verify that the headers were flattened correctly | D. Define a named range in Excel that excludes the title block and connect Tableau to that named range
Tableau expects column headers or row values in the first row, so introductory text and multi-level headers cause problems. Data Interpreter can remove title text and flatten hierarchical headers, and Tableau says to verify that the headers were flattened correctly; if the title cannot be removed from the source, a named range that excludes it is another documented option. Tableau does not ignore title rows automatically, and stacked headers are not turned into a hierarchy; they must be flattened. Tableau Bridge provides connectivity for private data to Tableau Cloud and does not clean files.

Question 5

Cloud Kicks keeps warehouse data in a niche analytical database that isn't listed under Connect in Tableau Desktop. The vendor supplies a 64-bit ODBC driver that implements the ODBC standard, and there is no connector on Tableau Exchange. Which connection method should the Tableau consultant recommend?
A. The Other Databases (JDBC) connector, with the vendor's ODBC driver copied into the Drivers folder
B. A custom SQL query entered against a Microsoft SQL Server connection that points at the database host
C. An Excel export of the warehouse tables each morning, connected with the Microsoft Excel connector
D. The Other Databases (ODBC) connector with the vendor's ODBC driver, using a DSN for portability

Accepted Answer

Correct answer: D. The Other Databases (ODBC) connector with the vendor's ODBC driver, using a DSN for portability
When a database isn't listed under Connect, help says you might connect with the Other Databases (ODBC) connector if the database driver implements the ODBC standard, and it recommends DSNs for publishing and cross-platform portability (a named connector, when one exists, should always be preferred). The Other Databases (JDBC) connector needs a driver that implements the JDBC standard, supplied as a .jar file, so it cannot load an ODBC driver. A SQL Server connection can't reach a different database engine, and daily Excel exports add manual steps and lose currency.

Question 6

An AW Computing extract uses Aggregate data for visible dimensions. The visible dimensions are Region, Product and Order Date (no roll-up); Order ID was hidden, and Sales has its default aggregation of SUM. The sales director now wants the average order value for each region. What should the Tableau consultant tell the director?
A. The extract can't support it, because each row now holds summed sales for a region, product and day, so the orders are gone
B. Change the aggregation of Sales to AVG in the view, because the aggregated extract kept the order-level average on every row
C. Divide by COUNTD([Order ID]), because fields that are hidden before extraction remain available to calculations in the extract
D. Add Order ID to the view as a dimension, because an aggregated extract restores hidden fields when a view requests them

Accepted Answer

Correct answer: A. The extract can't support it, because each row now holds summed sales for a region, product and day, so the orders are gone
Aggregate data for visible dimensions aggregates the measures with their default aggregation, here SUM, for each combination of the visible dimensions, and hidden fields are excluded from the extract. Each row therefore holds total sales for a region, product and day, so AVG of Sales returns the average of those daily totals rather than of orders. Order ID is not in the extract at all, so it can't be counted or added to the view. The requirement needs data at order grain (or an order count captured before aggregation), so the extract must be rebuilt.

Question 7

A Tableau consultant at Get Cloudy Consulting is rebuilding a data source in Tableau Desktop 2024.2. The billing dashboard must include only engagements that have at least one invoice, and the old data source achieved this with an inner join between Engagements and Invoices. In the new version the two tables are related instead, and unbilled engagements now appear in several views. What should the consultant recommend?
A. Change the relationship's cardinality to One to One so that every engagement must be matched by exactly one invoice row
B. Add an LOD expression that deduplicates the invoice amounts, because the relationship is multiplying the engagement rows
C. Set the relationship's referential integrity to All Values Match so that Tableau drops the engagements that have no invoices
D. Model the inner join as a physical join inside one logical table, because the purpose of the join is to filter the data

Accepted Answer

Correct answer: D. Model the inner join as a physical join inside one logical table, because the purpose of the join is to filter the data
Help explains that relationships work hard to recover unmatched values, sometimes using a left join where an inner join was intended to filter data, and that it can make sense to model that inner join as a physical join within a logical table. Referential integrity and cardinality are performance options that describe the data; setting All Values Match when unmatched values exist does not reliably filter them and can produce incorrect results. The problem is extra unmatched engagements rather than duplicated measures, and relationships already aggregate at each table's native level of detail, so a deduplicating LOD expression doesn't address it.

Question 8

A Northern Trail Outfitters scatter plot in Tableau Desktop shows one mark per transaction for about two million rows, and it is very slow, even though the analysis only needs one mark per store. The Analysis menu shows that measures are not aggregated. What should the Tableau consultant recommend?
A. Convert every measure to a dimension so that Tableau draws one mark for each unique value instead of each row
B. Create a FIXED LOD expression for each measure so that the values are computed in Tableau instead of the database
C. Turn on Analysis > Aggregate Measures and keep Store as the level of detail so the view returns one row per store
D. Change the extract to sample 10 percent of the rows so that the view draws fewer marks without changing the result

Accepted Answer

Correct answer: C. Turn on Analysis > Aggregate Measures and keep Store as the level of detail so the view returns one row per store
Help advises that if views are slow you should make sure you're working with aggregated measures, because disaggregated data can mean viewing many rows at once; selecting Analysis > Aggregate Measures reduces the rows returned to the level of detail of the view, here one per store. Treating measures as dimensions still returns one mark per distinct value. LOD expressions are computed in the database as nested selects and add work rather than remove it. Sampling changes the results instead of summarizing them correctly.

Question 9

A Northern Trail Outfitters calculation assigns loyalty tiers on a large extract and appears in several slow views. Refer to the calculation below:

IF [Points] >= 5000 THEN "Gold"
ELSE IF [Points] >= 2000 THEN "Silver"
ELSE IF [Points] >= 500 THEN "Bronze"
ELSE "Member"
END END END

Which change should the Tableau consultant recommend to make the logic more efficient?
A. Move the Member test to the top so the statement begins with the ELSE branch first
B. Rewrite the tests as CASE [Points] WHEN >= 5000 THEN "Gold" and so on for each tier
C. Rewrite the tests with ELSEIF in a single IF statement that ends with one END
D. Wrap each comparison in a FIXED expression so every tier is tested once per member

Accepted Answer

Correct answer: C. Rewrite the tests with ELSEIF in a single IF statement that ends with one END
Create Efficient Calculations explains that ELSEIF is faster than ELSE IF, because a nested IF computes a second IF statement rather than being evaluated as part of the first, so one IF with ELSEIF branches and a single END is the efficient form. CASE can be faster than IF for complex logic, but CASE compares a value to WHEN values and does not accept range comparisons such as >= 5000. Wrapping comparisons in FIXED expressions adds subqueries. An IF statement cannot start with its ELSE branch.

Question 10

A Cloud Kicks author created [Margin %] as SUM([Profit]) / SUM([Sales]) and wants it to appear as an average by default whenever it is dragged into a view. In the Data pane, Default Properties > Aggregation is not available for this field. What explains this?
A. The field must be published to Tableau Server before its default properties can be changed
B. Default aggregations are available only for extracts, so the data source must be extracted first
C. Default aggregations can be set only for dimensions, so the field must be converted to a dimension
D. A default aggregation cannot be set for a calculated field that already contains an aggregation

Accepted Answer

Correct answer: D. A default aggregation cannot be set for a calculated field that already contains an aggregation
Data Aggregation in Tableau states that you can set the default aggregation for any measure that isn't a calculated field that itself contains an aggregation, such as AVG([Discount]); [Margin %] is already an aggregate calculation, so Tableau wraps it in AGG and its aggregation is fixed by the formula. Default aggregation is a property of measures, not something that requires converting to a dimension. Neither an extract nor publishing is a prerequisite for setting default properties.

Question 11

Cumulus Financial has MONTH([Deposit Date]) on Rows and YEAR([Deposit Date]) on Columns for four years. The branch director wants each cell to show the deposits for the same month one year earlier. Which calculation and Compute Using setting should the Tableau consultant use?
A. LOOKUP(SUM([Deposits]), -1) computed using Table (down)
B. LOOKUP(SUM([Deposits]), FIRST()) computed using Table (across)
C. LOOKUP(SUM([Deposits]), -1) computed using Table (across)
D. LOOKUP(SUM([Deposits]), -12) computed using Table (down)

Accepted Answer

Correct answer: C. LOOKUP(SUM([Deposits]), -1) computed using Table (across)
Computing Table (across) addresses the years in each month row, so an offset of -1 returns the same month in the previous year column. Computed Table (down), an offset of -1 returns the previous month in the same year instead. An offset of -12 computed down points outside the 12 months of each year partition, and LOOKUP returns NULL when the target row cannot be determined. An offset of FIRST() always returns the first year in the row, not the prior year.

Question 12

At Universal Containers, Tableau Server views load slowly during peak hours. Monitoring shows high CPU and memory in the VizQL Server process when many users are active, and users do not need up-to-the-minute data. Which three actions should the Tableau consultant recommend? Choose 3 answers.
A. Set the data cache refresh frequency to always, so stale query results do not consume memory
B. Keep the data cache refresh frequency at low, so cached data is reused as much as possible
C. Add VizQL Server processes one at a time and measure the effect after each change
D. Increase the number of Cache Server processes, which the VizQL Server process queries frequently
E. Add Backgrounder processes on the same node, so they precompute views for the VizQL Server process

Accepted Answer

Correct answer: B. Keep the data cache refresh frequency at low, so cached data is reused as much as possible | C. Add VizQL Server processes one at a time and measure the effect after each change | D. Increase the number of Cache Server processes, which the VizQL Server process queries frequently
Optimize for User Traffic guidance says adjusting the number of VizQL Server processes is the most effective step, one process at a time with monitoring, that the VizQL Server process makes frequent requests to the Cache Server so more cache server processes can help, and that when users do not need the latest data the server should cache and reuse data as much as possible with tsm data-access caching. Setting caching to always makes every page load query the data again. Backgrounder processes compete with VizQL Server for CPU, so the guidance is to reduce them or move them to a dedicated node.

Question 13

Bloomington Caregivers' board receives a quarterly dashboard. An analyst proposes a Sankey diagram to show how admissions from two referral sources end in two discharge outcomes, a total of four flows. Several board members say they find complex charts hard to read. What should the Tableau consultant recommend?
A. A radar chart with one axis per referral source, so each discharge outcome is drawn as its own polygon over the same axes
B. The Sankey diagram, because any data with a before state and an after state should be shown as flows between the two states
C. A chord diagram, because its circular layout is easier for a non-technical audience to read than straight bars or lines
D. A simple bar chart of discharge outcome by referral source, because four flows for a non-technical audience do not need a Sankey

Accepted Answer

Correct answer: D. A simple bar chart of discharge outcome by referral source, because four flows for a non-technical audience do not need a Sankey
Sankey guidance says the chart works best with multiple inputs and outputs or complex flows, that a small data set is often better served by a simpler chart such as a bar chart, and that audiences who are not comfortable with complex charts may read a simpler chart more easily. Having a before and after state does not by itself justify a Sankey. A chord diagram is another complex relationship chart, so it does not help a less data-fluent board. A radar chart compares many variables per group and would turn four numbers into hard-to-compare polygons.

Question 14

Cloud Kicks wants a field that returns, for each customer, the share of that customer's sales that comes from their largest product category, so analysts can segment concentrated buyers in any view without the value changing as dimensions are added or removed. Which calculation should the Tableau consultant recommend?
A. {FIXED [Customer ID] : MAX({FIXED [Category] : SUM([Sales])})} / {FIXED [Customer ID] : SUM([Sales])}
B. {FIXED [Customer ID] : MAX({FIXED [Customer ID] : SUM([Sales])})} / {FIXED [Customer ID] : SUM([Sales])}
C. MAX({INCLUDE [Category] : SUM([Sales])}) / SUM([Sales]), computed with Customer ID in the view
D. {FIXED [Customer ID] : MAX({FIXED [Customer ID], [Category] : SUM([Sales])})} / {FIXED [Customer ID] : SUM([Sales])}

Accepted Answer

Correct answer: D. {FIXED [Customer ID] : MAX({FIXED [Customer ID], [Category] : SUM([Sales])})} / {FIXED [Customer ID] : SUM([Sales])}
The inner FIXED expression on Customer ID and Category returns each customer's total for each category, the outer FIXED expression on Customer ID takes the largest of those totals, and dividing by the customer's total sales gives the share; an LOD aggregate expression can contain other LOD expressions, and FIXED expressions compute without reference to the dimensions in the view, so the value stays the same in any view. An inner FIXED on Category alone uses only the dimension it names, so it returns each category's total across all customers. Leaving Category out of the inner declaration makes the inner total the customer's whole total, so that ratio is always 1. The INCLUDE version is correct only while Customer ID is the view's level of detail and changes when other dimensions are added.

Question 15

Cumulus Financial's dashboard has a Product Type list with a string parameter control; a calculated field uses the parameter to show the products of the chosen type. New product types are added every quarter and the typed parameter list is always out of date. Advisors also dislike switching to the control. What should the Tableau consultant recommend?
A. Replace the parameter with a context filter on Product Type, so the products are computed after each selection
B. Add a Product Type sheet and a parameter action that sets the parameter when an advisor clicks a product type
C. Rebuild the parameter as a fixed list every quarter and add a URL action that reloads the dashboard after edits
D. Replace the calculation with a set action, because parameters cannot be used in calculations that return dimensions

Accepted Answer

Correct answer: B. Add a Product Type sheet and a parameter action that sets the parameter when an advisor clicks a product type
With a parameter action, users change the parameter by clicking marks, which keeps them in the flow of the dashboard, and new product types that appear in the data are automatically shown and available to the parameter. A context filter changes the order of filtering but does not provide the click-driven selection. Rebuilding the list every quarter keeps the maintenance problem. Parameters can be used in calculated fields that return dimension values, so a set action is not required.

Question 16

A data steward at Universal Containers has the Creator site role on Tableau Server and wants to certify the published data sources in the Finance project. She is not a site administrator, and the company does not have Data Management. What should the Tableau consultant recommend?
A. Buy Data Management for the deployment, because every type of data label, including certification, requires it
B. Change the steward's site role to Explorer (can publish), because certification is limited to that site role
C. Make the steward the owner or a project leader of the Finance project, which lets her certify its data sources
D. Grant the steward the Overwrite capability on each Finance data source so that she can add the certified label

Accepted Answer

Correct answer: C. Make the steward the owner or a project leader of the Finance project, which lets her certify its data sources
To certify a published data source, a user must be a server or site administrator, or have the Explorer (can publish) or Creator site role and be the project owner or have the Project Leader capability for the project that holds the data source. Certifying published data sources is the one data label operation that does not require Data Management; it is needed only for certifying virtual connections, databases and tables. Overwrite is the capability for setting data quality warnings, not certification. Creators already qualify, so a site role change is unnecessary.

Question 17

Northern Trail Outfitters started with a centralized governance model in which a BI team builds every data source and dashboard. After a year of training, leadership wants to move to a delegated model as described in Tableau Blueprint. Which two characteristics should the Tableau consultant describe for the delegated model? Choose 2 answers.
A. IT remains the only producer of data sources and dashboards, which business users consume in a one-to-many manner as Viewers
B. Validation, promotion and certification are well understood by users of every skill level, and the boundaries between roles are fluid
C. Site administrators and data stewards are identified outside of IT, and some of them may have direct access to sources of data
D. Content authors use certified published data sources, and some consumers can web author and save derivative content to sandbox projects
E. Each department receives its own Tableau site, so that it can set its own governance policies without asking IT for changes

Accepted Answer

Correct answer: C. Site administrators and data stewards are identified outside of IT, and some of them may have direct access to sources of data | D. Content authors use certified published data sources, and some consumers can web author and save derivative content to sandbox projects
In Blueprint's delegated model, new roles such as site administrators and data stewards are introduced outside IT, authors build on certified published data sources, and some consumers get web authoring to save derivative content to sandbox projects, while validation, promotion and certification processes are still limited. IT producing all content for Viewers describes the centralized model. Fully understood processes and fluid roles describe the self-governing model. Per-department sites are not part of any Blueprint governance model and would block content sharing and promotion.

Question 18

Bloomington Caregivers' Tableau Cloud site administrator wants a monthly list of licensed users who have not signed in for more than 90 days, so that their licenses can be reassigned. Which approach should the Tableau consultant recommend?
A. Use the TS Users data source in Admin Insights, or the Starter workbook's site activity dashboard, with a 90-day threshold
B. Query the readonly user's tables in the Tableau repository and filter users by their last sign-in date
C. Open the Background Tasks for Non Extracts admin view and filter it to the sign-in tasks of the past 90 days
D. Ask each project leader to check Who Has Seen This View on their workbooks and report users who are missing

Accepted Answer

Correct answer: A. Use the TS Users data source in Admin Insights, or the Starter workbook's site activity dashboard, with a 90-day threshold
Admin Insights is the Tableau Cloud project with curated data sources; TS Users includes user, site role and activity data, and the Admin Insights Starter's site activity dashboard has an inactivity threshold parameter for finding users who have not signed in, accessed or published content. Repository access exists only on Tableau Server. Background Tasks for Non Extracts covers background jobs, not user sign-ins. Who Has Seen This View covers one view's viewers and cannot reveal users who never sign in.

Free Salesforce Tableau Consultant
Practice Questions

What is the Tableau Consultant exam like?

Questions by exam section

Evaluate Current State

Question 1 · Harder · Choose 2

Question 2 · Medium

Question 3 · Easier

Question 4 · Medium · Choose 2

Plan and Prepare Data Connections

Question 5 · Easier

Question 6 · Harder

Question 7 · Harder

Question 8 · Easier

Design and Troubleshoot Calculations and Workbooks

Question 9 · Medium

Question 10 · Medium

Question 11 · Medium

Question 12 · Harder · Choose 3

Question 13 · Easier

Question 14 · Harder

Question 15 · Medium

Establish Governance and Support Published Content

Question 16 · Easier

Question 17 · Medium · Choose 2

Question 18 · Easier

Where to go after these 18 questions

Get Tableau Consultant exam updates by email

Ready for a full Tableau Consultant practice exam?

Free Salesforce Tableau ConsultantPractice Questions