Why not play with each one and decide for yourself? R support is already built into SQL Server.

I got the chunks apart... So I guess it's a place to start.  You need Jeff's string splitter function for this, which is here.  And here's my code......

Got Google?
http://sqljason.com/2010/12/subscripts-and-superscripts-in-ssrs.html

Got it to work... probably a kludge, but ...Summarized in T-SQL first:
SELECT SalesRep
    , SUM(Sales2016) AS Total2016
    , SUM(Sales2017) AS Total2017
    , SUM(Sales2017) - SUM(Sales2016) AS Delta

That's a drilldown report. It's covered in Brian Larson's book... and probably all over the place.  Here's a super quick video on YouTube.

Okay, just tried it with my own dataset.  Turns out, I can't find a way to sort on an aggregated column in a matrix. 

So if you base your...

Don't think you need a new dataset. What you're describing sounds like you need a Matrix on your report.
Group on [Diagnosis Code], then at the intersection, do COUNT().
You...

Got Google?

http://sqlmag.com/t-sql/group-all

Did you look up the differences between DECIMAL, MONEY, and SMALLMONEY?

Did you look up the difference between SMALLMONEY and MONEY datatypes? MONEY should work fine.

Don't use BETWEEN with Times.  use <. > 
SELECT *
FROM MyTable
WHERE EventDate < @SomeTime AND EventDate>=DATEADD(minute,-10,@SomeTime)

What do you consider "unique messages"?  Do you mean something like "common sentences"?  I would probably use Jeff's splitter function and write the results to at least a temporary table...

SELECT TabA.Col1, TablA.Col2, TabA.Col3, TabB.X, TabB.Y
FROM TabA CROSS JOIN TabB

What do you mean by "missing days"?