Add a Bubble Layer to the Power BI Azure Maps Visual - Sizing Points That Actually Read

Put a hundred stores on a map and you have a scatter of identical dots that tells you where things are but nothing about how they are doing. Make those dots grow and shrink with revenue, and colour them by whether they hit target, and suddenly the same map tells you your biggest branch is also your worst performer, sitting there as a large red circle in the middle of Melbourne. That second map is what a bubble layer gives you, and it is one of the more genuinely useful things in the Azure Maps visual.

A bubble layer plots each location as a circle, and it lets you drive two extra dimensions off your data: the size of the bubble and its colour. So a single map can carry location, magnitude, and category all at once. For a country as spread out as Australia, where so much analysis comes down to "what is happening in each region," that is a lot of information in one glance.

I have built these for retail, logistics, and property clients, and the bubble layer is usually where a flat point map turns into something a manager will actually use. Microsoft's documentation on adding a bubble layer covers the settings. Here is the practical version, including the bits that go wrong.

What the bubble layer gives you

The Azure Maps visual in Power BI supports several layer types, and the bubble layer is the default one that appears when you drop location data in. On its own it just draws a dot per location. The value comes from two fields you can add.

Size. Drop a measure into the Size well and each bubble scales with that value. Revenue, order count, headcount, tonnes shipped, whatever your magnitude is. Bigger number, bigger bubble. This is the single change that turns a location map into a performance map.

Legend (colour). Drop a category into the Legend well and bubbles are coloured by it. Region, product line, status, tier. Now you can see clustering, so if all your red bubbles are on the east coast and the blue ones are in the west, that pattern shows up immediately.

You can also control the range of bubble sizes, the transparency, the outline, and the zoom behaviour. Transparency matters more than people expect, and I will get to why.

Getting the location right first

None of the bubble styling helps if the bubbles are in the wrong place. Everything I have written before about geocoding applies here. If you are plotting Australian locations by name or postcode, give the visual latitude and longitude wherever you can, and set the data categories on your location fields properly. A bubble sized by revenue is worse than useless if it is sitting over the wrong city, because now you have a big confident circle pointing at the wrong answer.

For any map that is going to inform a decision, get real coordinates into your model during data preparation and plot those. The bubble layer plots exactly what you give it, so give it coordinates you have checked. This is data modelling work as much as visual work, which is why it comes up constantly in our Power BI consulting engagements and usually gets fixed in the data layer, not the report.

Where bubbles work well

Regional performance at a glance. Sales by store, sized by revenue, coloured by target status. One map, and a regional manager can see who is big, who is small, and who is in trouble. The large red bubble is the conversation you need to have this week.

Distribution and coverage. Plotting where your customers are, sized by spend, shows you concentration. A logistics client of ours used this to see that a huge slice of their volume came from a tight cluster in western Sydney, which changed where they put a depot. A table would never have made that jump off the page the way the map did.

Comparing magnitude across a spread-out country. Australia's population sits in a handful of coastal clusters with vast empty space between. Bubbles handle that well because the size does the talking. A single big bubble over Perth reads clearly even though it is a thousand-odd kilometres from the next one.

Where it falls apart

Now the honest bit, because the bubble layer has real failure modes.

Overlap in dense areas. This is the big one. Put a lot of bubbles in a small area, like every store across metropolitan Sydney, and the big ones swallow the small ones. You end up with a blob where individual points are invisible. Transparency helps a little, because overlapping semi-transparent bubbles show through each other, but it only goes so far. If your data is dense in cities and sparse in the bush, the map is readable in the outback and a mess in the metro, which is usually the opposite of what you want. Sometimes the honest answer is that a map is the wrong visual for dense urban data and a table or a suburb-level aggregation reads better.

Size perception is unreliable. People are bad at judging area. A bubble that is twice the value does not look twice as big to the eye, it looks a bit bigger, and a bubble ten times the value looks maybe four times as big. So bubble size communicates rough magnitude and rank, not precise comparison. If someone needs to know that Brisbane did exactly 1.7 times Adelaide, do not make them read it off bubble area. Put the number in a tooltip or a companion table. Bubbles are for "this one is clearly bigger," not for measurement.

Outliers wreck the scale. One enormous value and every other bubble shrinks to a pinprick to fit it in. A single flagship store doing ten times everyone else will flatten the rest of your map into dots. Watch your size range, and consider capping or handling outliers so the bulk of your data stays legible. A map where 95 per cent of the bubbles are the same tiny size is not showing you anything.

Too many colours in the legend. Same rule as any visual. If your Legend field has fifteen categories, you get fifteen colours nobody can distinguish, especially on small overlapping bubbles. Keep the colour dimension to a handful of meaningful categories. Status with three states works. Product with twenty does not.

Practical settings that matter

Turn transparency up a bit. Fully opaque bubbles hide everything behind them. Semi-transparent bubbles let overlapping points show through and make dense areas at least partly readable. It is a small setting with a big effect.

Set a sensible minimum and maximum bubble size. The defaults are often too wide, so small values vanish and large ones dominate. Tightening the range keeps the whole dataset visible.

Use the tooltip properly. Because bubble size is imprecise, the tooltip is where the real numbers live. Load it with the exact figures, the location name, and any context the reader needs when they hover on a bubble that caught their eye. The map draws attention, the tooltip delivers the detail.

Think about zoom. Auto-zoom can frame your data well or badly depending on outliers. A single mis-geocoded point in the North Atlantic will zoom the whole map out to include it, and suddenly Australia is a tiny cluster in the corner. Which, again, is why the location data has to be clean first.

My rule of thumb

Use a bubble layer when your data is spread out enough that size reads clearly, and when rough magnitude plus one category is genuinely what the reader needs. Retail branches across the country, customers by region, volume by depot, these are the natural fits. If your points are crammed into a few city blocks, or the reader needs precise comparison, reach for a different visual or aggregate up a level.

Done well it is one of the fastest ways to turn regional data into a decision. Done carelessly it is a pretty picture that hides more than it shows.

Mapping like this usually sits inside a wider business intelligence build, and the map is only ever as trustworthy as the Microsoft Fabric model feeding it. If you want Power BI maps built by people who get the geography and the modelling right the first time, take a look at our services or get in touch.