Dashboards are where an analyst’s work becomes visible to managers. This guide covers everything a beginner needs to build trustworthy Power BI reports: preparing data with Power Query, building a clean data model, writing DAX measures, time intelligence, filters and row-level security, then choosing charts, designing a clear dashboard and presenting it in an interview. A section on Tableau covers the equivalent skills for that tool.
If you are new to data, start with our Excel guide and SQL guide first. Power BI is much easier when you already think in tables, filters and joins.
Contents
- 1. Power BI vs Tableau: which to learn first
- 2. Power Query: merge and append
- 3. Relationships and the star schema
- 4. Date tables
- 5. DAX basics: measures vs calculated columns
- 6. CALCULATE and filter context
- 7. SUMX and iterator functions
- 8. Time intelligence: YTD, MTD and previous year
- 9. Filters: visual, page and report level
- 10. Row-level security
- 11. Choosing the right chart
- 12. Dashboard design mistakes to avoid
- 13. Tableau calculated fields
- 14. Tableau dashboard actions and filters
- 15. Presenting a dashboard in an interview
1. Power BI vs Tableau: which to learn first
Power BI and Tableau are the two dashboard tools that appear most in analyst job posts. Both connect to Excel files and databases, build charts, filters and dashboards, and share reports with other people.
- Power BI is made by Microsoft and fits closely with Excel and other Microsoft tools, so many companies that already use Microsoft products choose it. Power Query (data preparation) and DAX (formulas) are its central skills.
- Tableau is known for flexible, polished visuals and drag-and-drop exploration, and is common in companies with dedicated analytics teams.
To decide, check five job posts for the city and role you want. If most mention Power BI, start there. The core ideas, such as clean data, good chart choice and clear titles, carry over to the other tool.
2. Power Query: merge and append
Data rarely comes in one file. Monthly sales arrive as separate sheets, and customer details live in another system. Power Query, available in both Power BI and Excel, combines them with two operations beginners often confuse. (For Power Query basics such as changing types and removing duplicates, see the Excel guide.)
Append: stack tables for more rows
Use Append when tables have the same columns, such as January, February and March sales files: Home, then Append Queries. Columns are matched by name, not position, so “Amount” and “amount ” become two separate columns; clean and rename first. For many files in a folder, use Get Data from Folder and Combine Files, so new monthly files are picked up automatically on refresh.
Merge: join tables for more columns
Use Merge to add columns from a related table, such as customer city on each order: Home, then Merge Queries, select the matching column in both tables, choose a join kind, then expand the new column.
- Left outer: all rows from the first table, with matched details where found. The most common choice.
- Inner: only rows that match in both.
- Left anti: rows in the first table with no match, ideal for finding orders with missing customers.
- Full outer: everything from both tables.
Watch for mismatched data types (text “101” vs number 101), extra spaces or different capitalisation in keys, and duplicate keys in the lookup table, which multiply rows. Check row counts before and after. Often a model relationship (next section) is simpler and lighter than a merge.
3. Relationships and the star schema
When a visual shows the same number in every row, or a slicer seems to do nothing, the cause is almost always the data model, not the chart. Relationships tell Power BI how tables connect and which way filters travel.
One-to-many and the star schema
A good model has a fact table of events, such as Sales, at the centre, connected to dimension tables such as Customers, Products and Date. Each customer appears once in Customers but many times in Sales: a one-to-many relationship, shown as 1 on the dimension side and * on the fact side. Drawn out, it looks like a star. Power BI needs the “one” side to have unique values; if it refuses to create the relationship, your dimension has duplicate keys to clean.
Filter direction
By default, filters flow from the one side to the many side: picking a city in a Customers slicer filters Sales. This single direction is what you want most of the time. Bi-directional filtering can feel convenient but often creates ambiguous paths, slower reports and totals that are hard to explain.
Common mistakes
- Accidental many-to-many: joining two fact tables directly instead of through a shared dimension.
- Inactive relationships (dotted lines), which are only used when a measure calls USERELATIONSHIP.
- Mismatched data types between key columns.
- No proper date table (see the next section).
Before building visuals, open the model view and check that every fact table connects to its dimensions with one-to-many, single-direction lines.
4. Date tables
A date table has one row per day, with columns such as year, month, quarter and weekday. Without one, measures that compare periods can return wrong totals or blanks.
- Create a table with one row for every day in the range your data covers, with no gaps.
- Add columns for year, month name, quarter and day of week.
- Mark it as a date table in Power BI so time intelligence functions can use it.
- Create a one-to-many relationship from the date table to your fact table’s date column.
Check the result by comparing one month’s total in the report with a simple SUM on the source data.
5. DAX basics: measures vs calculated columns
DAX is Power BI’s formula language. The first idea to master is the difference between a calculated column and a measure.
- Calculated column: adds a value to every row, worked out once when the data loads. Use it when you need to filter or group by the value, such as an age band for each customer.
- Measure: calculated when a visual needs it, based on the filters in place. Use it for totals, averages and ratios that should change with the report filters, such as total sales or profit margin.
A simple rule: if you will put it on an axis or in a slicer, it is probably a column; if you will show it as a number that changes with filters, it is a measure. Measures to learn first: SUM and COUNTROWS for totals, DIVIDE for safe ratios, CALCULATE to change filters, and time intelligence functions.
6. CALCULATE and filter context
CALCULATE is the most important function in DAX. It evaluates an expression under a different set of filters, which is how one measure can give “sales for the North region” or “sales last year”.
North Sales = CALCULATE(SUM(Sales[Amount]), Regions[Name] = "North")Every visual starts with filters from slicers, rows and columns: its filter context. CALCULATE adds to or replaces those filters. When a result looks wrong, check which filters are active on that visual. To make a measure ignore a slicer, for example to show a share of the total, use CALCULATE with ALL or REMOVEFILTERS on that column.
7. SUMX and iterator functions
SUM adds up a column. SUMX goes through a table row by row, calculates an expression for each row, and then adds the results.
Total Sales = SUM(Sales[Amount])
Total Revenue = SUMX(Sales, Sales[Quantity] * Sales[Price])
Gross Margin = SUMX(Sales, Sales[Quantity] * (Sales[Price] - RELATED(Products[Cost])))A plain SUM cannot calculate revenue here, because the multiplication must happen per row first. Inside SUMX, DAX works in row context: it knows the current row’s values. That is different from filter context, and understanding both is the core of DAX. RELATED follows a many-to-one relationship so each sales row can read its product’s cost.
Other iterators include AVERAGEX, MAXX, MINX, COUNTX and FILTER (often used inside CALCULATE). Iterators run over every row, so prefer a plain SUM when no row-level logic is needed. Common mistakes: using SUM on a column that needs multiplying first, using RELATED without a relationship, and nesting iterators unnecessarily.
8. Time intelligence: YTD, MTD and previous year
Managers rarely ask “what were sales?” They ask “how are we doing compared with last year?” Time intelligence functions answer this, provided you have a marked date table (section 4).
Sales YTD = TOTALYTD([Total Sales], 'Date'[Date])
Sales MTD = TOTALMTD([Total Sales], 'Date'[Date])
Sales PY = CALCULATE([Total Sales], SAMEPERIODLASTYEAR('Date'[Date]))
Sales YoY % = DIVIDE([Total Sales] - [Sales PY], [Sales PY])For an Indian financial year starting in April, add the year-end date: TOTALYTD([Total Sales], 'Date'[Date], "31/3"). DIVIDE avoids errors when last year’s value is zero. For a rolling three months, use DATESINPERIOD inside CALCULATE.
Common mistakes: using the sales table’s date column instead of the date table, not marking the date table, showing YoY % without the actual values, and comparing a partial current month with a full previous month.
9. Filters: visual, page and report level
“Why does this card show a different total from that chart?” The answer is almost always filters, which Power BI applies at several levels at once:
- Visual-level filters apply to one visual only.
- Page-level filters apply to every visual on one page.
- Report-level filters apply to every page.
- Slicers are visible on the page so users can change them.
On top of these, clicking a bar cross-filters or cross-highlights other visuals. Use slicers for the few choices users change often, such as year or region, and the Filters pane for fixed rules such as excluding test data, locking or hiding those filters. Use Format, then Edit interactions, to stop a company-total card reacting to a regional chart, and View, then Sync slicers, so a year chosen on one page applies to all pages.
To keep reports trustworthy, show the active selections on the page, keep hidden filters to a minimum, and test by clearing all slicers and checking totals against the source data.
10. Row-level security
Row-level security (RLS) shows different rows of the same report to different people. A regional manager sees only their region while head office sees everything, without separate copies of the report.
- Create a role in Power BI Desktop, such as “North Manager”.
- Add a filter on the table, for example Region equals North.
- Publish the report, then assign people to each role in the Power BI service.
- Test each role with “View as” before sharing.
RLS limits what a viewer sees in the report but is not a replacement for database permissions. Test every role, because a missing filter can show all rows to the wrong person, and keep the rule simple: one filter on one column is easiest to check.
11. Choosing the right chart
The right chart makes a finding obvious; the wrong one hides it. Start with the question, then choose the chart that answers it most directly:
- Comparing categories: a bar chart, sorted from largest to smallest.
- Change over time: a line chart.
- Part of a whole: a stacked bar, or a pie chart only when there are very few parts.
- Relationship between two numbers: a scatter plot.
- One key number: a single large figure with a comparison next to it.
Avoid 3D charts, which distort values, and bar charts whose axis does not start at zero. Write titles that state the finding: “South region sales grew fastest this quarter” rather than “Sales by region”.
12. Dashboard design mistakes to avoid
A dashboard should answer one question quickly. Beginners often build a page full of charts that look impressive but hide the answer.
- Too many charts: five or six is usually enough for one audience.
- No title that states the point: “Monthly revenue fell 8% in October” beats “Revenue”.
- Too many colours: use one colour for the item you want people to notice.
- 3D or pie charts for comparisons: a bar chart is easier to read.
- No filters: people want to see their own region or month.
A simple test: show the dashboard to someone who did not build it and ask for the main message after 10 seconds. If they cannot say it, simplify the page.
13. Tableau calculated fields
A calculated field creates a new value from existing columns, turning raw sales and cost into profit, margin or a late-delivery flag.
- Right-click a field in the Data pane, or choose Create Calculated Field.
- Give it a clear name, such as Profit Margin.
- Write the formula, for example
SUM([Profit]) / SUM([Sales]). - Click OK and drag the new field onto the view.
Check that the result is right for one row you can verify by hand, that the aggregation matches the chart (a sum of ratios gives the wrong answer), and that the name is clear to the next reader.
14. Tableau dashboard actions and filters
An interactive Tableau dashboard lets a manager click a region and see every chart update. Filters and dashboard actions make that possible.
- Filters: drag a field to the Filters shelf, choose Show Filter, then use Apply to Worksheets so one filter controls several charts.
- Filter actions: Dashboard, then Actions, then Add Action, then Filter. Set a source sheet, such as a map of regions, and target sheets. Clicking Karnataka on the map then filters the trend and product table. “Show all values” when cleared is usually the friendliest setting.
- Highlight actions: keep all data visible but grey out everything except the selection, useful when context matters.
- Parameters: let users switch what a chart shows, such as Sales, Profit or Orders, by using the parameter in a calculated field with CASE.
Keep dashboards fast and clear: use extracts for large data, limit quick filters, add a short instruction such as “Click a region to filter”, and fix the dashboard size. A good portfolio piece is a sales dashboard with a region map, a monthly trend and a top-products table connected by one filter action and one parameter.
15. Presenting a dashboard in an interview
Interviewers often ask you to explain a dashboard you built. Use this four-part structure:
- The business question, in one sentence: “This dashboard shows which products drive revenue in each region.”
- The data: where it came from, how many rows it had, and the one cleaning step that mattered most.
- The finding: one or two insights, each with the number behind it. Do not walk through every chart.
- The next step: what you would do next, such as testing a discount or checking a second region.
Record yourself explaining one dashboard in two minutes, listen back and cut anything that does not help the interviewer understand the result. For a complete project to present, follow SQL to Power BI: a small sales dashboard project, and for the latest features see what changed in Power BI in October 2026.
Learn Power BI with ScholarLogic
Power BI, DAX and Tableau are taught with real projects in the 5-month Data Analytics course in Bangalore at ScholarLogic, BTM Layout. See what you learn in the Power BI course, or book a free demo class and a counsellor will share the fee and the next batch dates. To book, call 78923 93639.
