ELI5: What's the difference between a metric, a dimension, and a filter?

Okay, buckle up, because I see people in the Ideogram dashboards—and honestly in AWS Cost Explorer, GCP Billing Reports, you name it—mixing these three up constantly, and it’s making their cost analysis about as useful as a screen door on a submarine. Let's break down the holy trinity of cost data slicing: metrics, dimensions, and filters.

Think of your raw cloud billing data as a giant, unorganized spreadsheet with a million rows. It's just a firehose of "someone spent something, somewhere."

* **A Metric** is the **number you actually care about**. It's the quantitative measure, the thing you sum or average. In cost contexts, this is almost always a dollar amount (your `UnblendedCost` or `UsageQuantity`). It's the "what" you're measuring. In a chart, this is typically on the Y-axis.
* **A Dimension** is the **category or column you use to group that metric**. It's the "by" in "show me my cost *by* service" or "*by* region" or "*by* project." Dimensions are the attributes attached to every line item—like `ServiceName`, `Region`, `UsageType`, `ProjectId`. You use a dimension to split your metric into meaningful buckets. In a chart, this often ends up on the X-axis or as a legend.
* **A Filter** is a **static constraint you apply to the entire query**. It's you saying "only show me data **where** this dimension equals (or contains) this specific value." It reduces the overall dataset you're looking at before any grouping happens. "Show me my S3 costs" means you're adding a filter: `WHERE ServiceName = 'Amazon S3'`.

Here's a pseudo-code example that would make a billing report:

```sql
SELECT
SUM(UnblendedCost) as TotalCost,
ServiceName
FROM billing_data
WHERE Region = 'us-east-1'
AND Date BETWEEN '2024-10-01' AND '2024-10-31'
GROUP BY ServiceName;
```

**The classic screw-up:** People use a dimension as a filter and then wonder why their totals don't make sense. If you want to know what percentage of your bill is S3, you group **by** `ServiceName` and look for the S3 slice. If you only ever **filter** *to* S3, you'll never see that it's 60% of your spend because you've blinded yourself to the other 40%.

In Ideogram's visualization tools, you'll typically:
1. Choose your **Metric** (Cost, Usage).
2. Choose your primary **Dimension** to break it down (Service, Account, Environment).
3. Apply **Filters** to zoom in on a specific slice of time, a specific project, or to exclude test resources.

Master this, and you can actually start hunting down those pesky dev environments running `m6g.16xlarge` instances for a forgotten Jupyter notebook. Your cloud bill is too high.