Power BI - How to Show the Data Behind a Visual and Why It Matters

The single biggest cause of trust problems with a Power BI report is not a bad chart, a missing filter, or even slow performance. It is the moment a senior leader looks at a number and says "where did that come from?", and nobody in the room can answer them in under thirty seconds.

We see this play out at almost every Power BI engagement we run. The CFO is staring at a sales by region bar chart, asks why Queensland is down 12%, and the report author has to go back to the desktop file, write a DAX query, paste it into the wrong workspace, and then quietly say "I'll get back to you tomorrow." Trust dies a little bit every time that happens.

The fix is built into Power BI and has been for years. It is the "Show as a table" feature (and its newer sibling, "Visual table"). Most analysts I meet either do not know it exists, or use it badly enough that they have written it off. So let's go through how it actually works, where it bites, and the workflow we now teach every client team.

Two related features, slightly different jobs

Power BI gives you two ways to see the data behind a visual, and the distinction matters more than the documentation suggests.

Visual table (or "Show as a table" in the service) takes the entire visual and lays out every row of the underlying aggregated data next to it. If your bar chart shows ten product categories, you get a table with ten rows. This is the right tool when someone challenges the chart as a whole.

Data point table (or "Show data point as a table" in the service) is different. You right-click on a single bar or slice, and Power BI shows you every detail row that fed into that one number. If a category bar shows $1.2 million in sales, this gives you the actual transactions or rows aggregated up to that figure. This is the right tool when someone challenges a single number.

In a board meeting, you use Data point table. In a model review, you use Visual table. Get these two confused and you waste a lot of time scrolling.

How to actually trigger them

For Visual table in Desktop, select a visual, head to the Data/Drill tab on the ribbon, and hit Visual table. The visual stays in place and the data shows below or beside it depending on your orientation. You can toggle between horizontal and vertical with the little icon in the top right.

For Data point table, you need to be hovering over a specific data point. Right-click on the bar or slice and choose Show data point as a table. There is also a button in the Data/Drill tab, but you still have to click a data point after pressing it.

In the Power BI Service it is roughly the same flow, just hidden behind the More options (...) menu in the top corner of each visual. The right-click option only appears when you are hovering directly over a data element, which is genuinely a bit fiddly on dense charts.

Where this saves us in real client work

Here are three patterns where this feature has paid for itself many times over.

The reconciliation conversation. A logistics client we worked with had a recurring weekly meeting where the operations team would dispute the numbers in the executive dashboard. The data team would spend a day each week building reconciliation queries. We turned on Show as a table for the contested visuals, added a CSV export button right next to them, and the disputes mostly stopped within a month. When people can see the exact rows feeding a number, they argue with the data less and the source system more, which is the right place for that argument.

The DAX debugging trick. When a calculated measure looks wrong, the fastest debugging path is often: drop the measure into a visual, drop in the dimensions you suspect, hit Visual table, and read the actual numbers. You will catch context filter problems and incorrect totals far quicker than reading the DAX. Performance Analyzer has its place, but for "why is this number weird" questions, Visual table is faster.

The handover trick. When we hand a model over to an internal team, we walk them through Show as a table on every key visual. It does two things. It proves the visual is consistent with the data they already trust in Excel, and it teaches them how to answer the "where did that come from" question themselves, without paging us.

If you want help setting up that kind of trustworthy reporting layer, our Power BI consultants team does this kind of model and report review work as part of every engagement.

The limitations that will trip you up

Microsoft buries the limitations near the bottom of the docs, but they matter a lot when you are designing a report.

Show as a table does not work on every visual type. It is unsupported for Card, KPI, Key influencers, Q&A, Smart narrative, Metrics app visuals, paginated report visuals, Power Apps visuals, ArcGIS maps, and Power Automate visuals. If you build a dashboard that is mostly cards and KPIs, you have just designed yourself out of the easiest "show me the data" experience your users have. We now actively avoid card-heavy designs in any report that needs to survive executive scrutiny.

Data point table is even more limited. If you have a measure in the Value field of a visual, Data point table is grayed out. This is the one that catches most people, because every well-built Power BI report uses measures. If you want users to be able to drill into a single data point, you sometimes need to build a separate detail page that is bookmark-linked from the main visual, rather than relying on Data point table. It is annoying, but it is the truth.

Neither feature works in PDF or static PowerPoint exports. If your executive workflow ends with a PDF being emailed to the board, the data tables are not coming with the chart. You need to plan for that, usually by exporting to an Excel workbook in parallel or designing a dedicated data page for the export pack.

Live multidimensional models break it. If you are connecting Power BI live to a multidimensional SSAS model, the data view features either do not work or break in unhelpful ways. This was a real problem for a client we helped migrate off a legacy Analysis Services Multidimensional cube. We could not give business users the underlying-data experience until we moved to a tabular semantic model. If you are still on multidimensional and want this kind of self-serve capability, plan a migration.

What we now bake into our standard report builds

A few small habits have made a big difference for our clients.

We turn on tooltips that include the row count being aggregated, so users get a hint of the underlying volume before they even need to hit Show as a table. It is a one-time setup that pays off forever.

We design at least one detail page per report, accessible via a drillthrough button, that shows the underlying rows of any selected data point. This is our fallback for the cases where Data point table is grayed out because of measures. The drillthrough page is the real workhorse for executive trust.

We train users to export to Excel rather than CSV when they want to share the data. The Excel export includes formatting and a hyperlink back to the source visual, which is a small thing that helps with audit trails when the spreadsheet ends up in an email thread three weeks later.

For our enterprise clients with Microsoft Fabric consultants engagements, we go further and tie this back to lineage in the Fabric workspace, so a curious executive can click from a number on a dashboard all the way back to the source table in the lakehouse. That level of traceability is the difference between people trusting the report and people quietly rebuilding the report in Excel on their laptop.

The short version

Show as a table and Show data point as a table are old, slightly hidden features in Power BI that solve one of the biggest cultural problems with self-serve BI, which is trust. They are not perfect. The limitations around measures and multidimensional models are real and you need to design around them. But for any report that matters, you should be turning them on, training users how to find them, and building drillthrough fallbacks for the cases where they do not work.

If you want a second pair of eyes on your reports, or you are running into the trust problem at executive level, this is the kind of thing we do every week. Drop us a line through our contact page and we can take a look.

Original Microsoft reference: Show the data used to create a Power BI visual