How to Use Filters in Power BI: Visual, Page and Report Level Filters with Top N and Advanced Filtering

By Ravindra Bagale · Last updated: 16 October 2026

Practice files: blinkit-sales-2025-small.csv (the 12 orders on this page), blinkit-sales-2025.csv (360 orders, 16 products, for the practice task) and blinkit-product-hierarchy.csv (for section 7).

Made-up data: fictional Blinkit-style orders from Maharashtra cities in 2025, for practice only. Amounts are in ₹ and dates are DD-MM-YYYY.

1. The problem

Your manager asks three questions:

1. Which 3 categories bring the most revenue?
2. Which are our top 5 products?
3. How much do we sell of anything called biscuit?

A plain chart of revenue by product shows every product side by side, and the numbers on the bars read "1K" and "0.64K". You need to cut the chart down to what was asked, and show numbers people can read.

2. The idea in simple words

A Power BI Desktop file is a report. A report has pages (the tabs at the bottom), and each page holds visuals. (Dashboards are built in the Power BI Service by pinning visuals from reports; a one-page report is often loosely called a dashboard.)

The Filters pane has three levels, from narrow to wide:

Two filter types do most of the work:

• Top N keeps the top (or bottom) N items by a value, such as Top 3 categories by revenue. It works like Excel's TAKE on a sorted list.
• Advanced filtering keeps items that match a rule, such as product contains "biscuit".

3. The data

The 12 orders in blinkit-sales-2025-small.csv (columns: order_id, order_date, product_id, product, category, sub_category, city, sales, revenue):

Total revenue ₹3,850. By category: Dairy & Bakery 1,600, Snacks & Branded Foods 880, Sweets 840, Staples 530.

By product, top 6: ghee 1 l 1,300, kaju katli 640, toor dal 320, shrikhand 300, bhakarwadi 240, jowar atta 210.

4. Step by step

A new page and a chart that is too big

1. Load blinkit-sales-2025-small.csv (Get data › Text/CSV › Load). Set revenue to a whole number if needed.
2. At the bottom, click + to add a report page. Double-click the tab and name it Filters.
3. Insert a Clustered column chart. X-axis: product. Y-axis: revenue. You get 11 bars (one product was not sold); with the 360-row file you get 16, and a real shop gives hundreds.

Top 5 products (visual level)

4. Click the chart so it is selected. Open the Filters pane. Under Filters on this visual, expand product.
5. Filter type: Top N. Show items: Top, 5. By value: drag revenue into the box. Click Apply filter.
6. The chart now shows 5 bars: ghee 1 l, kaju katli 250 g, toor dal 1 kg, shrikhand 500 g, bhakarwadi 200 g.

Top 3 categories

7. Add a second clustered column chart: X-axis category, Y-axis revenue.
8. With it selected, Filters on this visual › category › Top N › Top 3 › By value revenue › Apply filter. Staples (530) disappears.

Products that contain a word (Advanced filtering)

9. Add a third chart (product × revenue). Filters on this visual › product › Filter type Advanced filtering.
10. Show items when the value: contains, type biscuit. Click Apply filter. Two bars: jeera biscuits 250 g (200) and masala chai biscuit 200 g (90).

Page and report level

11. Drag city from the Data pane to Filters on this page and tick Pune. Every chart on this page now shows Pune only. Untick it again.
12. Drag category to Filters on all pages to filter every page at once. Remove it when you are done (hover › ✕).

Readable numbers

13. Select the Top 5 chart. Open the Format pane (the paintbrush) › Visual › Data labels › On.
14. Under Values: Display units None, Value decimal places 0, font Segoe UI, size 14, bold, and a dark colour.

5. Your result should look like this

With the page filter city = Pune, the category chart shows Sweets 840 · Dairy & Bakery 650 · Snacks & Branded Foods 290: Pune revenue is ₹1,780.

Data labels read 1,300, not 1.3K.

6. What happened

• Top N ranked the products by the sum of revenue and kept the first 5. The 6th (jowar atta, 210) dropped out.
• The Top 3 category filter worked on category totals, not on single orders.
• "contains biscuit" matched both products whose name has the letters biscuit. It did not match butter cookies, even though it sits in the Biscuits & Cookies sub-category: a text filter only reads the field you filter.
• The visual filters belong to one chart each. The page filter on city changed all three charts; the report filter would also change the other pages.
• Display units None and 0 decimals show the full value instead of a rounded K value.

7. Explore further

• Ties in Top N. If two items tie at the cut-off, Power BI keeps both, so "Top 5" can show 6 bars. In the 360-row practice file there are no ties in the top 5.
• Top N needs a "by value". "Top 5" alone means nothing; Top 5 by revenue and Top 5 by units sold can be different products.
• "biscuit" vs "biscuits". contains biscuits matches only jeera biscuits; the singular biscuit catches both. Type the shortest word that is common to all the names you want.
• Basic filtering is the tick-box list. Use it when you know the exact items (Pune, Nashik); use Advanced when you need a rule (contains, starts with, is not blank).
• Filters pane vs Format pane. The Filters pane decides which data a visual shows. The Format pane (paintbrush) decides how it looks: labels, colours, fonts.

When every bar shows the same number. Load blinkit-product-hierarchy.csv as a second table, producthierarchy. Build a chart with category from producthierarchy and revenue from sales. If the two tables have no relationship (for example because you turned off relationship autodetect), every bar shows 3,850, the grand total: Power BI cannot tell which sale belongs to which category. Open Model view: there is no line between the tables. Drag product_id from producthierarchy onto product_id in sales (one-to-many). Back in Report view the bars become 1,600 · 880 · 840 · 530. You could also merge the two tables in Power Query, as in How to Merge Tables in Power BI Power Query (Left Outer vs Inner Join).

8. Common mistakes

9. Practise on your own

10. Think like a professional

• Business view. A report answers questions. Top N and page filters turn a crowded chart into one answer per visual.
• Make filters visible. Users cannot see visual filters unless they open the pane. Put the rule in the title: "Top 5 products by revenue".
• Report filters are powerful and easy to forget. A forgotten "Filters on all pages" entry quietly changes every number. Check it before you publish.
• Relationships first. If bars look identical, check Model view before you blame the data.

Interview questions

1. What are the three filter levels in Power BI? Visual (this visual), page (this page) and report (all pages).
2. How do you show the top 5 products by revenue? Filters on this visual › product › Top N › Top 5 › By value revenue › Apply filter.
3. Why does every bar in a chart show the same total? The category and the value come from tables with no relationship.
4. Basic vs advanced filtering? Basic is a list of ticked items; advanced uses rules like contains or is not blank.

Recap

Ravi Sir's Cards for this lesson

Revision card

मराठी cards