Tableau Analytics-DA-201 Data Analyst Study Guide

What Does the Tableau Analytics-DA-201 Exam Cover?

Salesforce Analytics-DA-201, the Tableau Data Analyst certification, is a 60-question, 105-minute exam requiring 65 percent to pass, priced at $200 USD with a $100 retake. It covers four weighted domains: Explore and Analyze Data (41%), Create Content (26%), Connect to and Transform Data (24%), and Publish and Manage Content (9%).

The 41% domain to prioritise

The 41 percent concentration is the single most useful planning fact. Roughly 25 questions live in calculations, table calculations, LOD expressions, filters, parameters, sets, and analytics features – more than double the next domain. Publishing at 9 percent is around five questions and does not warrant proportional study.

Pacing and question style

At 105 seconds per question the pace is comfortable, which suits an exam that frequently describes a required outcome and asks which calculation or feature produces it. The Tableau Desktop help documentation is the authoritative reference throughout.

How Are Data Connections and Transformations Tested?

Connect to and Transform Data is worth 24 percent, covering connections to files, databases, and Tableau Server or Cloud, plus preparation and cleaning, Tableau Prep transformations, and field property customisation. It tests getting data into a usable shape.

Live versus extract

The live versus extract decision is the foundational connection topic. A live connection queries the source on every interaction, so data is current but performance depends on the source system. An extract is a snapshot stored in Tableau’s own optimised format, which is faster and works offline but is only as fresh as its last refresh.

Choosing by situation

The examinable reasoning is situational: real-time dashboards need live connections, while large historical datasets on a slow source benefit enormously from extracts. Questions describe requirements around freshness and performance and expect you to choose.

Field properties

Field properties are the detail-oriented half of the domain. Data types, default aggregations, geographic roles, and aliases all affect how a field behaves everywhere it is used. Aliases in particular are worth understanding – they change displayed values without altering the underlying data, which is the correct answer whenever a scenario wants cleaner labels rather than transformed data.

Tableau Prep

Tableau Prep handles the heavier transformation work: cleaning steps, aggregations, pivots, and output to an extract or database. The examinable boundary is when preparation belongs in Prep versus in the data source versus in calculated fields – Prep suits reshaping that would otherwise be repeated across many workbooks. The Tableau Prep documentation covers the transformation types.

Joins, Unions, Relationships, and Blends: What Is the Difference?

These four combining methods are among the most reliably examined topics in the connect domain, and candidates confuse them constantly. Each combines data differently and suits different situations.

• Union – stacks rows from tables with the same structure; more rows, same columns
• Join – matches rows on a key and combines columns; row count depends on join type and can duplicate
• Relationship – a flexible logical link where Tableau chooses the join at query time based on the fields used
• Blend – combines data from separate sources at the visualisation level, aggregating each independently

Relationships

Relationships are the modern default and the concept most worth understanding properly. Unlike a fixed join, a relationship does not flatten the data up front – Tableau determines the appropriate join for each visualisation depending on which fields are used. That avoids the classic problem where joining at the wrong granularity duplicates rows and inflates every measure.

Blending

Blending remains relevant for genuinely separate data sources that cannot be joined, such as a database and a spreadsheet on different systems. Its examinable property is that it aggregates each source independently before combining, which is why blended results sometimes differ from what a join would produce.

The typical question pattern

The question pattern is consistent: a scenario describes two datasets and a requirement, and you identify the correct combining method. Duplicate rows after combining almost always point at a join at the wrong granularity, and the fix is usually a relationship.

Which Calculation Types Must You Know Cold?

Calculated fields sit at the heart of the 41 percent domain and cover date, string, logical, number, and aggregate functions. Fluency here is the single highest-value preparation investment for this exam.

Logical calculations

Logical calculations are the most frequently examined. Know the difference between IF, which evaluates conditions in sequence, and CASE, which matches a single expression against values – and that IIF is the compact form for simple two-outcome conditions. Questions present a required categorisation and expect the correct construct.

Date functions

Date functions matter because time is in almost every analysis. DATEPART extracts a component such as month or year, DATEDIFF measures an interval between two dates, DATEADD shifts a date, and DATETRUNC rounds a date down to a specified level. That last one is the most useful and most under-known – truncating to month is how you group daily transactions into monthly totals correctly.

Row-level versus aggregate

The aggregate distinction is the conceptual trap. A row-level calculation is evaluated for every underlying row before aggregation, while an aggregate calculation operates on already-aggregated values. SUM([Profit])/SUM([Sales]) and SUM([Profit]/[Sales]) return different answers, and understanding why is genuinely examinable – the first divides totals, the second averages ratios.

String functions

String functions round out the set: SPLIT, CONTAINS, LEFT, RIGHT, MID, REPLACE, and TRIM. These appear in data cleaning scenarios where the required transformation is small enough not to warrant Prep.

Why Do Table Calculations Trip Candidates Up?

Table calculations compute on the aggregated results already in the visualisation rather than on the underlying data. They cover running totals, moving averages, rankings, percentiles, and percent-of-total – and their difficulty lies entirely in addressing and partitioning.

Partitioning and addressing

Partitioning defines the groups within which the calculation restarts; addressing defines the direction it moves. A running total partitioned by region and addressed along date restarts at each region and accumulates through time. Change the partition and the same calculation answers a different question.

Why they feel unpredictable

That dependency is the reason table calculations feel unpredictable to candidates who have only used the defaults. The calculation is correct; the addressing describes something other than what was intended. Exam questions frequently present a result that looks wrong and expect you to identify the partitioning as the cause.

The common functions

Know the common functions and what each needs. RUNNING_SUM accumulates along the addressing direction. WINDOW_AVG averages across a defined window. RANK orders within the partition. LOOKUP retrieves a value from an offset position, which is how period-over-period comparisons are built. INDEX and SIZE support position-aware logic.

They run after aggregation

The critical property to carry into the exam is that table calculations operate after aggregation and after most filters – which means they see only what is already in the view. A table calculation cannot reach data that a filter removed, and that constraint answers a category of question about unexpected results.

What Do Level of Detail Expressions Actually Do?

Level of detail expressions compute aggregations at a granularity different from the visualisation. Named explicitly in the syllabus, they are the most powerful and most misunderstood feature in the exam.

FIXED, INCLUDE, and EXCLUDE

The three types answer different questions. FIXED computes at the specified dimensions regardless of what is in the view – it ignores the visualisation’s granularity entirely. INCLUDE adds dimensions to whatever the view already has, computing at a finer level then aggregating up. EXCLUDE removes dimensions from the view’s granularity, computing at a coarser level.

Master FIXED first

FIXED is the one to master first because it solves the most common real problem: comparing a value against a total that should not change as the user filters or drills. A customer’s lifetime spend, computed FIXED at customer level, stays correct whatever dimensions appear in the view.

FIXED and the order of operations

The property that generates exam questions is FIXED’s position in the order of operations. It computes before dimension filters are applied, which means a FIXED expression ignores those filters unless they are promoted to context filters. That is exactly the behaviour candidates find surprising, and exactly why it is tested.

INCLUDE and EXCLUDE

INCLUDE and EXCLUDE are more situational. INCLUDE suits computing an average of per-customer values shown at regional level; EXCLUDE suits showing a value against a regional total while the view is broken down further. The LOD expressions documentation works through each with examples.

How Does Tableau’s Order of Operations Affect Filters?

Tableau applies filters in a defined sequence, and understanding that order explains most otherwise-baffling results. It is not a named syllabus item, but it underlies a substantial share of the analysis domain’s questions.

The filter sequence

The sequence runs from extract filters through data source filters, context filters, dimension filters, FIXED LOD calculations, measure filters, and finally table calculation filters. Each stage sees only what earlier stages passed through.

Two practical consequences

The practically important consequences are two. A FIXED LOD expression is unaffected by a normal dimension filter, because the LOD computes first – promoting that filter to a context filter is what makes it apply. And a table calculation filter hides values without recomputing the calculation, which is how you show a top-N list whose percentages still reflect the full dataset.

Sets and groups

Sets and groups interact with this too. A set is a dynamic or fixed subset that can be used as a filter or a dimension; a group is a static combination of members into a single value. Bins convert continuous measures into discrete ranges, which is what makes histograms possible.

Parameters

Parameters complete the picture as user-controlled inputs. A parameter alone does nothing – it must be referenced by a calculation, filter, or reference line to have effect. Questions about interactivity that does not work frequently trace to a parameter that was created but never wired to anything.

What Does the Create Content Domain Require?

Create Content is worth 26 percent, covering chart construction, dashboards and stories, formatting, and interactivity through actions and dynamic visibility. It tests building things that communicate rather than things that merely render.

Choosing the right chart

Chart type selection follows conventional analytical reasoning: line charts for trends over time, bar charts for categorical comparison, scatter plots for relationships between measures, maps for geographic distribution, and tables when precise values matter more than shape. Questions describe what a user needs to see and expect the appropriate visual.

Dual-axis and combined charts

Dual-axis and combined charts are the construction technique most often examined, along with the distinction between a synchronised axis – where both measures share a scale and are genuinely comparable – and independent axes, which can mislead by implying a relationship that the scaling manufactured.

Dashboard actions

Dashboard actions are the interactivity mechanism. Filter actions use a selection in one view to filter another, highlight actions emphasise related marks without filtering, and URL actions open external resources. Knowing which action produces a described behaviour is a reliable question type.

Dynamic visibility

Dynamic visibility is the newer capability worth knowing: dashboard objects can be shown or hidden based on a parameter or calculation, which allows a single dashboard to serve multiple audiences rather than duplicating it. Layout containers and device-specific layouts round out the formatting material.

How Is Publishing and Content Management Examined?

Publish and Manage Content is the smallest domain at 9 percent – around five questions – covering publishing workbooks and data sources, extract refresh scheduling, alerts, subscriptions, and permissions. Calibrate your effort accordingly.

Published data sources

The published data source concept is the most valuable idea here. Publishing a data source separately from workbooks means connection details, calculations, and field formatting are defined once and reused, so a change propagates everywhere rather than needing repetition in every workbook.

Extract refresh scheduling

Extract refresh scheduling is the operational topic. Understand the difference between a full refresh, which replaces the whole extract, and an incremental refresh, which appends new rows only – incremental is far faster but does not capture changes to existing rows, which is precisely the trade-off a question would test.

Permissions

Permissions determine who can view, interact with, edit, and download content, and are set at project, workbook, or view level with inheritance from the project down. The examinable principle is that permissions should generally be managed at project level rather than item by item.

Subscriptions and alerts

Subscriptions and alerts close the domain: subscriptions deliver content on a schedule, while data-driven alerts notify when a value crosses a threshold. The publishing documentation covers both.

The boundary with senior credentials

This domain also marks the boundary with the more senior Tableau credentials. Analysts who go on to own platform architecture rather than content typically move toward the Tableau certified architect path, where server topology and governance replace calculations as the main subject.

How Should You Prepare for Analytics-DA-201?

Six to eight weeks at six hours per week suits analysts with Tableau experience. Effective preparation weights time toward the 41 percent analysis domain and is done in Tableau rather than in documentation – calculations are learned by writing them and getting them wrong.

A career decision point

Analytics-DA-201 also sits at a useful career decision point. It validates hands-on analysis, whereas the advisory track covered in the Tableau consultant career guide shifts toward client engagement and solution design – worth knowing which direction you are heading before choosing what to study next.

1. Weeks one to two – connections and combining. Practise live versus extract decisions, then combine the same datasets with a join, a relationship, and a blend to see the different results. Deliberately join at the wrong granularity and watch measures inflate.
2. Weeks three to five – analysis. Half your total time belongs here. Write calculations of every type, drill the row-level versus aggregate distinction, then work table calculations by changing partitioning until the results stop surprising you.
3. Week six – LOD and order of operations. Build FIXED, INCLUDE, and EXCLUDE expressions for the same question. Apply a dimension filter to a FIXED calculation, observe that nothing changes, then promote it to a context filter.
4. Week seven – content creation. Build dual-axis charts, dashboard actions, and dynamic visibility. Construct one dashboard that serves two audiences through parameter-driven visibility.
5. Week eight – publishing and review. Cover published data sources, refresh types, and permissions briefly, then move to timed practice.

Predict before you run

The habit that matters most is predicting a result before running it. Say what a table calculation will return given its partitioning, then check – that loop builds the intuition the exam tests far faster than reading about addressing. Timed work through the Analytics-DA-201 practice exam questions confirms whether your time is landing where the weightings are.

Frequently Asked Questions

How many questions are on the Analytics-DA-201 exam?

The exam contains 60 questions to be completed in 105 minutes, allowing roughly 105 seconds per question. The pace suits its scenario-based style.

What is the passing score for Analytics-DA-201?

You need 65 percent, meaning 39 of the 60 questions correct. The retake fee is $100 USD, lower than the $200 initial registration.

Which domain carries the most weight?

Explore and Analyze Data at 41 percent, roughly 25 questions. It covers calculations, table calculations, LOD expressions, filters, parameters, sets, and analytics features.

What is the difference between a join and a relationship?

A join combines tables at a fixed granularity when the data source is built. A relationship is a logical link where Tableau chooses the appropriate join per visualisation, which avoids duplicating rows when granularities differ.

What does a FIXED LOD expression do?

It computes an aggregation at specified dimensions regardless of the visualisation’s granularity. Critically, it evaluates before dimension filters, so those filters do not affect it unless promoted to context filters.

Why do table calculations produce unexpected results?

Almost always because of partitioning and addressing. The calculation is usually correct but is restarting or moving along a different grouping than intended. Changing the partition changes what question the calculation answers.

When should you use an extract instead of a live connection?

When source performance is poor, the dataset is large and historical, or offline access is needed. Live connections suit requirements for current data where the source can handle the query load.

What is the difference between full and incremental extract refresh?

A full refresh replaces the entire extract. An incremental refresh appends new rows only, which is much faster but does not capture changes to existing rows.

Why does a parameter appear to do nothing?

Because a parameter is only an input control. It has no effect until referenced by a calculated field, filter, reference line, or dynamic visibility rule.

How long should I study for Analytics-DA-201?

Six to eight weeks at around six hours per week suits analysts with Tableau experience. Give the analysis domain roughly half your total time, matching its 41 percent weighting.

Conclusion

Analytics-DA-201 is a calculation exam wearing a visualisation label. Explore and Analyze Data carries 41 percent, and preparation that spends its time building dashboards is preparing for the 26 percent domain while neglecting the largest one.

Three concepts account for a disproportionate share of the difficulty. Table calculation partitioning explains most unexpected results. FIXED LOD expressions and their position before dimension filters explain most of the rest. And the row-level versus aggregate distinction quietly determines whether a calculation is right at all.

Prepare by predicting before you run. State what a calculation will return, then check it – and when you are wrong, work out why. That loop is what converts Tableau familiarity into the analytical fluency this exam is actually measuring.