Word for consolidating a dataset to a single type

How about exchanged?

Graph 1 represents each series in the local currency.

Graph 2 represents each currency exchanged to [reporting currency here] using the [datasource] rates as of [datestamp].

Maybe you are thinking of standardized, to bring to a standard measurement.

You used it in your title. I like:

The second graph is consolidated to the same currency.

As with any business artifact, its creation is driven by the intended audience and context. (Is it to the Board of Directors? Is it coming from accountants? Is it to be a standard internal report that has cross-dependent use as a data source? Etc.) That said, simplicity is best. Letting the graph speak for itself. Simply use, “In dollars” as a footnote descriptor to the graph. (Or “Euros” or whatever the audience expects.)

If within a paragraph you need to talk about the graph just say “the consolidated view below.” Where you get into trouble is when you manipulate source data vs. the representation. Then you might include a footer with explicit reference to conversion rate source and date (again depending on proper exactitude of context and use). You probably don’t need to give currency exchange rate source and date, but you do need to make sure the currency values are concurrent, and not – for example – 6 months apart.

“Normalized” is a reserved word in database circles for database design. “Massaged” data is used quite often but not about financial data. (E.g. J. Smith = John Smith, [or does it?]) “Massaged data” intuitively includes rounding errors and such, but it can be pejorative if not careful on a hard report. “Standardized” is rarely used, and can misrepresent – as if all local currencies are inferior to the Euro/Dollar/Yen standard, or whatever. So I would just use “the consolidated view below”. Financial statements talk of “consolidated” funds across global entities.

I am also a fan of normalized, but here is another option: When one set of data is translated into coordinate system, we sometimes say the data is projected. This is more commonly used when converting from a 3-dimensional system to a 2-dimensional system, such as using a flat map to represent a globe (consider the Mercator Projection)