Building a Power BI interactive dashboard, from data model to pinned tile

In Power BI, a dashboard is a single page of tiles that you pin in the Power BI service, and you can't slice or filter it directly. The clicking, filtering and drilling people expect from a Power BI interactive dashboard happen in the report underneath, or in a report page pinned as a live tile. So every build has two halves: an interactive report in Power BI Desktop, then a dashboard in the service that points to it.
The test for a finished build is plain. A sales leader sees the West region down 9% and asks why. She should find the cause herself in two or three clicks, without filing a ticket. Menu names below match Microsoft Learn as of October 2026.
The short version
- Build and model the report in Power BI Desktop, publish it, then pin visuals or a whole report page to a dashboard in the service.
- Model a star schema before the first visual. The model decides whether slicers feel instant or sluggish.
- Q&A goes away in February 2027. Copilot replaces it, but only on paid Fabric (F2 or higher) or Premium (P1 or higher) capacity.
- No new SSRS versions ship after SQL Server 2022, and SSRS 2022 is supported through January 11, 2033.
Before you start
You need Power BI Desktop, a free Windows app for connecting data, shaping it and building visuals. Microsoft recommends installing it from the Microsoft Store so it updates itself.
You also need a license that matches how you'll share. Microsoft lists three per-user licenses: Fabric (Free), Power BI Pro and Premium Per User (PPU). Free users can build reports and dashboards for their own use. Publishing to shared workspaces or collaborating takes Pro or PPU, unless the content sits in a workspace on Premium capacity or Fabric F64 and above. Dashboards follow the same line: creating one in a shared workspace needs Pro or PPU, while your own My Workspace doesn't.
On cost, Microsoft's Power BI pricing page listed Pro at $14.00 per user/month and PPU at $24.00 per user/month, paid yearly, when we checked on October 23, 2026. Actual prices vary by country and currency, and Fabric capacity is priced separately.
Building the Power BI interactive dashboard, step by step
Creating an interactive dashboard in Power BI goes faster in this order, because it saves rework. Each step leans on the one before it, so jumping straight to visuals usually means rebuilding them once the model changes.
1. Write down the three questions the page must answer
Start with the reader, not the data. Ask the people who'll use it which decisions they make every week and which numbers they argue about. "Which regions are behind plan, and which products are dragging them?" is a dashboard question. "Show all sales data" isn't.
Each question should map to one or two visuals. Anything that doesn't answer one belongs on a drillthrough page, or nowhere. These are common starting points, each with the interactive feature that tends to matter most:
| Team | What the dashboard shows | Feature that earns its place |
|---|---|---|
| Sales | Pipeline by stage, bookings vs target | Drillthrough to the account |
| Finance | Budget vs actual, variance by cost center | Drill-down from year to quarter to month |
| Operations | Inventory levels, late orders, supplier on-time rate | Data alerts on dashboard tiles |
| Leadership | 5–6 KPIs pulled from several reports | Subscriptions and a featured dashboard on Home |
| Customers or partners | Order or usage data inside your own portal | Embedded reports with row-level security |
If only three people read the numbers once a month and they already live in Excel, stop here. Not every team needs BI tools, and a well-built workbook may cost less to maintain.
2. Connect your data sources in Power BI Desktop
Open Power BI Desktop and select Home > Get data. Choose your source, pick tables in the Navigator pane and select Load. For an Excel workbook, it's Home > Get data > Excel.
Real dashboards rarely draw from one system. Expect an ERP, your CRM, a finance system and a spreadsheet or two. Power BI connects to cloud services and on-premises sources alike, though data behind your firewall needs an on-premises data gateway once you publish.
Some teams land app data in a spreadsheet or warehouse first, using a data integration tool. This comparison of Coupler.io and ten alternatives covers pricing, connectors and compliance.
3. Clean and shape the data in Power Query
Select Transform data to open the Power Query Editor. Remove columns you won't report on, filter out rows you don't need and set the right data type on every column. Then select Close & Apply.
It's the dull step that decides whether anyone trusts the result. If a sales total doesn't match the finance system, readers stop exploring and start emailing you. And if the source data can't be trusted at all yet, fix that first, because interactivity spreads bad numbers faster.
4. Model a star schema before the first visual
Open Model view and sort your tables into facts and dimensions. Microsoft's star schema guidance explains the split. Dimension tables (date, product, customer, region) handle filtering and grouping. Fact tables (sales orders, stock balances) handle summarizing. Relate them one-to-many, with the dimension on the "one" side.
Three modeling habits save trouble later. Build a proper date table rather than relying on auto date/time. Microsoft recommends turning the global option off when it isn't needed, because every date column otherwise gets a hidden table. Write explicit DAX measures for totals and ratios, so every visual calculates revenue the same way. And create hierarchies (Year > Quarter > Month, Category > Subcategory > Product) by right-clicking a column, since drill-down depends on them.
5. Lay out the report page
Switch to Report view, check fields in the Data pane and pick chart types in the Visualizations pane. A sensible first page:
- Cards across the top for the 3–4 headline numbers
- A line chart for trends over time
- A bar or column chart for comparing regions, products or channels
- A table or matrix for detail, filtered with Top N so it doesn't load every row
Put the number that matters most top left, where the eye lands first. Label in business language: "Net revenue, last 30 days" beats "Sum of Amt." Use pie and donut charts only for a part-to-whole split with few slices.
Keep one page for one audience. A CFO and a warehouse supervisor need different pages, even from the same semantic model. Go easy on large background images, which can block pinning, and test every custom visual for performance, as Microsoft advises.
6. Add slicers so readers can filter on the page
Microsoft describes slicers as on-canvas filters that keep the current filter state visible. You can choose the Slicer visual (list, tile, dropdown or date picker), the Button slicer, the List slicer or the Input slicer for typed values.
Two settings are worth knowing. Sync slicers, on the View ribbon, applies one slicer's selection across several report pages. Apply all and Clear all buttons, added from Insert > Buttons, let readers make several choices and apply them at once, which cuts unnecessary queries.
7. Decide how visuals cross-filter each other
By default, selecting a data point in one visual cross-filters or cross-highlights the rest of the page. Cross-filtering removes data that doesn't apply. Cross-highlighting keeps everything but dims what doesn't match.
To change it, select a visual and choose Format > Edit interactions. Icons appear on every other visual; pick Filter, Highlight or None for each. The settings save with the report, so authors and readers see the same behavior. Line charts, scatter charts and maps can only be cross-filtered.
Readers often don't know a bar is clickable. A short subtitle such as "Select a region to filter the page" does more than another visual would.
8. Add drill-down and drillthrough
The two names sound alike and do different jobs.
Drill-down moves through a hierarchy inside one visual, from year to quarter to month. Put a hierarchy on the axis and readers get drill icons in the visual header. They can drill up, turn on drill down, go to the next level or expand all down one level.
Drillthrough sends readers to a different page, filtered to what they clicked. To set it up:
- Add a new page and name it, for example "Customer details."
- In the Visualizations pane, find the Drillthrough section.
- Drag a field such as
Customer Nameinto the Drillthrough filters well. Power BI adds a back button for you. - Add visuals that make sense for a single customer.
Readers right-click a data point, choose Drillthrough and see only that item's data. It's how the West region question gets answered without cramming more visuals onto the main page.
9. Use report page tooltips for detail on hover
A report page tooltip is a small page that appears when a reader hovers over a data point. Create a page, set its canvas size to Tooltip under Canvas settings in the Format pane, turn on Tooltip in the Page information card and drag fields into Tooltip fields.
Use them sparingly. They aren't interactive, screen readers can't read them and dashboards don't support them.
10. Turn saved views into buttons with bookmarks
A bookmark saves a page's state: filters, slicers, sort order, drill location and which visuals are visible. On the View tab, open Bookmarks, arrange the page and select Add.
To make it clickable, choose Insert > Buttons > Blank, turn Action on in the Format button pane and set Type to Bookmark. A common use is a pair of buttons that swaps a chart for a table. Clear the bookmark's Data option for that, or switching views will reset the reader's filters.
Buttons also handle page navigation, drillthrough, web links and going back. Test them with Ctrl + click while editing.
11. Restrict rows with row-level security
If regional managers should see only their region, set up row-level security (RLS). In Power BI Desktop, select Modeling > Manage roles > New, name the role and write a DAX filter such as [Region] = "West". For one role that adapts to each person, use USERPRINCIPALNAME() against a user-mapping table.
RLS applies only to workspace Viewers. Admins, Members and Contributors see every row, which is why testing as yourself proves nothing.
12. Publish to the Power BI service and assign roles
Select Publish on the Home ribbon, sign in and choose a workspace. Per Microsoft's publishing guide, you'll get a semantic model and a report there. Open the semantic model's More options menu, select Security, add members to your RLS roles and check each with Test as role.
13. Pin visuals or a whole report page to a dashboard
Open the report in the service and select Edit. Hover over a visual, select the pin icon, then choose New dashboard and name it, or pick an Existing dashboard. Repeat for each tile.
Pinning single visuals gives you something close to snapshots, each linking back to its report. If readers should keep slicers and cross-filtering on the dashboard, pin the entire report page instead; those tiles stay live. With tiles in place, set data alerts on supported tiles and let people subscribe.
14. Add a mobile layout, then share
Plenty of readers will open the dashboard on a phone. In Desktop, mobile layout view can auto-create a phone version, which you then trim. It shows in the iOS and Android apps in portrait orientation, and it publishes with the main report.
Share the dashboard directly, or publish it in an app for wider audiences. Reports can also live where people already work: a Power BI tab in a Microsoft Teams channel, or a report web part on a SharePoint site page. The Teams tab embeds reports, not dashboards. Free-license readers need the workspace on Premium capacity or Fabric F64 and above; otherwise they need Pro or PPU.
To put the same content inside your own product, use Power BI embedded analytics, which places reports, dashboards and tiles in websites and web apps. "Embed for your customers" suits external users, who don't need a Power BI license, though production use needs a capacity. "Embed for your organization" suits internal portals, where each user signs in with a license.
Power BI dashboard vs report vs app: where the interactivity lives
People mix these up because all three look like canvases full of visuals. Microsoft's comparison draws the line clearly:
| Capability | Dashboard | Report |
|---|---|---|
| Pages | One page | One or more pages |
| Data sources | One or more reports and one or more semantic models | A single semantic model per report |
| Filtering and slicing | No (you can filter a tile in focus mode, but can't save it) | Yes: slicers, filters, cross-highlighting |
| Drilling down in visuals | Only if you pin an entire report page | Yes |
| Built in Power BI Desktop | No, Power BI service only | Yes |
| Data alerts | Yes, on some tiles | No |
| Subscriptions | Subscribe to the dashboard | Subscribe to a report page |
| See the model's tables and fields | No, but you can export data | Yes |
| Favorites and featuring on colleagues' Home page | Yes | Yes |
| Natural language queries (Q&A, until February 2027) | Yes | Yes, with edit permissions |
Source: Microsoft Learn, dashboards versus reports, checked October 24, 2026.
Treat the dashboard as the front door, pulling tiles from several reports, and the report as the place for exploring. Microsoft's optimization guide suggests dashboards as the first landing page for a practical reason: most tiles are served from a cache, so they load faster.
The app, now called a workspace app, adds distribution rather than interactivity. It packages dashboards, reports and other content from one workspace for a wider audience. Readers can filter, highlight and sort, but can't change anything. Each workspace gets one app, with up to 25 audience groups that each see only their own content. Creating an app takes Pro or PPU, and readers need Pro or PPU too, unless the workspace sits on Premium capacity or Fabric F64 or higher. If you add a dashboard but leave out the reports behind its tiles, those tiles show an error.
So build a report when people need to answer follow-up questions. Build a dashboard when leaders want one page of key performance indicators (KPIs) drawn from several reports. Wrap them in an app once more than a small team needs the content.
Q&A retires in February 2027, so plan around Copilot
Microsoft states that Q&A experiences are going away in February 2027 and recommends Copilot instead. Don't design new dashboards around the Q&A box or the Q&A visual.
Microsoft describes Copilot as an artificial intelligence (AI) assistant for insights and report creation. According to Microsoft's Copilot overview, business users can ask questions about a report or summarize it from the Copilot pane. Authors can use it to create and edit report pages, write DAX queries and add narrative summaries. So yes, Copilot can draft report pages you then pin to a dashboard. The report Copilot pane is generally available; the standalone, full-screen Copilot is in preview.
The conditions matter more than the demo. Copilot needs paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher), and trial capacities and free SKUs aren't supported. Microsoft says a Pro or PPU license alone isn't enough, and an admin has to turn on the tenant setting for Copilot and Azure OpenAI features.
Microsoft also warns that without a prepared semantic model, Copilot can misread the data and return generic or inaccurate results. In practice, Copilot answers well on a clean model with clear names and descriptions. On a messy one, it guesses.
Where Power BI sits in Microsoft business intelligence
Microsoft business intelligence is the set of Microsoft tools for collecting, modeling, analyzing and reporting business data. Power BI is now its center. Microsoft calls Power BI its business analytics platform for turning data into actionable insights. It's one of five products in Microsoft Power Platform and a core workload of Microsoft Fabric. Drag-and-drop report building means non-technical users can make simple reports while the data team owns the model.
Power BI business intelligence is a set of parts, each with its own job:
| Part | What it's for |
|---|---|
| Power BI Desktop | Free Windows app for data modeling and report creation |
| Power BI service | Cloud platform for publishing, sharing, dashboards, apps, refresh, alerts and security |
| Power BI Mobile apps | Viewing reports and dashboards on phones and tablets |
| Power BI Report Builder | Paginated reports: structured, printable reports such as invoices |
| Power BI Report Server | Keeping reports on-premises, with the option to move to the cloud later |
| Power BI embedded analytics | Putting reports, dashboards and tiles inside your own web apps |
Microsoft Fabric covers data integration, engineering, warehousing, real-time analytics and reporting, all sharing one logical data lake called OneLake. Power BI is the workload that turns that data into reports and dashboards, and existing Power BI content and workspaces stay as they are. Fabric touches dashboard projects in two ways. Microsoft is retiring Power BI Premium per-capacity SKUs in favor of Fabric F SKUs. And when Power Query struggles with large volumes, a Fabric warehouse or lakehouse is one place to build the data warehouse that the star schema guidance recommends. Our comparison of Microsoft Fabric vs Power BI goes further.
Power BI isn't the right home for every analytics job, either. Second-by-second streaming belongs in Fabric Real-Time Intelligence, which is built for data in motion. A data team that works mostly in notebooks may prefer Hex, which calls itself an analytics platform for humans and agents, with notebooks, data apps and dashboards in one workspace.
Moving from SSRS to Power BI after SQL Server 2025
A lot of Microsoft business intelligence still runs on SQL Server Reporting Services (SSRS). Microsoft's consolidation FAQ says that starting with SQL Server 2025, all on-premises reporting moves under Power BI Report Server (PBIRS). No new SSRS versions will ship. SSRS 2022 keeps security updates and support through January 11, 2033, so there's time to plan rather than rush.
Microsoft describes PBIRS as a superset of SSRS. It runs existing RDL (paginated) reports and adds interactive Power BI reports, data modeling and custom visuals. It comes with SQL Server 2025 Enterprise and Standard editions. On SQL Server 2022 and earlier, only Enterprise core licenses with active Software Assurance include it.
| Path | What it means | Fits when |
|---|---|---|
| Stay on SSRS 2022 | Supported through January 11, 2033, with no new versions | Stable operational reports and no change planned yet |
| Move to Power BI Report Server | Existing RDL reports are fully compatible; you add interactive reports on-premises | Data or reports must stay on-premises |
| Move to the Power BI service | RDL reports become paginated reports; exploration reports get rebuilt as interactive ones | You want dashboards, apps and cloud sharing |
Microsoft's plan for migrating .rdl reports to Power BI runs in four stages: before you start, pre-migration (discover, assess, prepare), migration and post-migration. Within that plan, four details prevent most of the pain:
- Inventory before you move anything. Microsoft suggests the free Azure Migrate tool to find report servers, plus a cleanup pass for reports nobody opens.
- Expect gaps. Mobile reports and KPIs can't be migrated, so rebuild them in Power BI Desktop with a mobile layout.
- Use the built-in tooling. Power BI Report Server and recent SSRS versions can publish .rdl files to Power BI directly; older versions use Microsoft's RDL Migration Tool on GitHub.
- Retest anything that filters on security fields. In the Power BI service, UserID returns the user principal name and ExecutionTime returns UTC.
Don't convert every report into an interactive one. Paginated reports are built for printing, PDFs and exact layouts, like a run of invoices that prints every row. Power BI reports are built for exploration. A useful filter: reports people export to Excel and then filter are your best candidates for an interactive dashboard. Our guide to Power BI paginated reports covers the fixed-layout side. For a larger estate, Brainvire's Microsoft BI team would start with that inventory, sorting each report into move, rebuild or retire before any migration tool runs.
Troubleshooting a slow or ignored dashboard
Most problems show up after launch, as either sluggish pages or empty usage stats. Start with the symptom:
| Symptom | Likely cause | Fix |
|---|---|---|
| The page takes seconds to respond to a slicer | Too many visuals, each sending its own query | Keep the visuals that answer your three questions; move detail to drillthrough pages and tooltips, as Microsoft highly recommends |
| One table visual drags the page down | An unfiltered table loading every row (Microsoft's example is 100 million rows) | Apply a Top N filter set well above what readers need |
| Filters are slow and DAX is hard to write | One flat, wide table holding everything | Split it into a star schema with a single date table |
| You can't tell which visual is slow | No measurement | On the Optimize ribbon, open Performance analyzer and select Start recording to see time spent on the DAX query, visual display and other work |
| Regional managers see other regions' data | RLS was tested by an admin, who sees every row | Use Test as role, then sign in as a real Viewer for dynamic rules |
| Nobody opens it after week two | A design problem, not a data problem | Watch two or three readers use it and remove whatever they ignore |
Several of these trace back to the model rather than the visuals. If you want outside help, our Power BI consulting team would look at the star schema, measures and RLS roles before touching a single chart. Brainvire is a Microsoft Solutions Partner, and when the dashboard is one piece of a wider data analytics program, the pipelines feeding it get the same review. Pick the dashboard your leadership argues about most, and tell us where it falls short.
FAQ
Can you create a dashboard in Power BI Desktop?
No. Dashboards exist only in the Power BI service. You build reports in Power BI Desktop, publish them, then pin visuals to a dashboard in the service. You can view dashboards in the Power BI mobile apps, but you can't create them on mobile devices.
Can a Power BI dashboard show real-time data?
Dashboard tiles update as new data reaches the underlying model, and pinned live report pages behave like reports. For DirectQuery models, Microsoft says the tile cache refreshes every hour by default, which you can change in the semantic model settings. For true streaming data, Fabric Real-Time Intelligence is the better fit.
Is Power BI hard to learn?
Building a basic report is approachable if you're comfortable with Excel. The harder skills are data modeling and DAX, which decide how fast and accurate an interactive report is. Microsoft Learn offers free tutorials, starting with getting started in Power BI Desktop.
Related reading
Embedded systems programming: how firmware is structured, written and testedJuly 28, 2021 · 12 min read
eLearning management systems: 18 LMS platforms grouped by who you trainDecember 16, 2024 · 13 min readPlanning a Power BI project in Southeast Asia? See Brainvire’s Power BI consulting in Singapore and Power BI consulting in Thailand.