This site [1] seems like a great complement to usafacts.org. It's great that these philanthropists are working on solutions to help us understand the empirical results of our policies and politicians.
Basically, it can look like spending on something has gone way up even if in reality it has gone down.
Doing the conversion is an elementary school thing. And not even that: just type "1000 dollars in 1947" to Google, and you'll get a number of links that do the conversion.
Sometimes it's the deceptive conversions that are a problem, so keeping in nominal sums helps with data integrity.
Firstly, they don't clearly explain that the figures aren't inflation-adjusted. You have to surmise that. And we're talking about line graphs. I think you are seriously overestimating the number of people who will think to do the conversion and follow through, rather than just looking at the general direction of the chart and a few figures.
I am not sure what you mean when you say that "sometimes deceptive conversions are the problem," but in my view there isn't much reason to produce a line graph for something like average salary over time or average social spending over time as a line graph without inflation-adjusted figures unless you either are ignorant of the issues with such data or intend to mislead people (and indeed such graphs are quite misleading since the numbers go up but not apace with inflation). I can't see much legitimate interest in such a graph.
I mean conversions where the index for conversion is chosen deceptively. The general inflation index is almost always at least tolerable, but occasionally the conversions use some other price index that produces more dramatic and politically desired results.