Double majoring only “helps” if the second major is in a complementary field that makes sense for the main field you’re entering. For example, a psychology student who double majors in statistics and applies for a quantitative psychology PhD program could have an advantage, since quant psych relies heavily on statistics. In this case, though, it’s not the double major in and of itself that does it - it’s the additional knowledge and coursework; it’d count the same if the person just took 10 classes in statistics but never declared a formal major. A psychology major who double majored in Spanish would derive no direct benefit in graduate admissions from the second major.
In your specific case, yes, the CMDA major could contribute a lot - but that’s because computational modeling is a huge part of statistics and advanced expertise in this area would be valued by a graduate program. But the mere fact of double majoring doesn’t make you look better, so if you could develop this expertise in some other way (like taking the three classes) then do that instead.
Also, don’t rely on the “I have heard.” I have seen so many students say this. There is a lot of information on data science out there right now, so I would do some Googling and especially look at some job ads for data scientists and see what languages they ask for (knowing that the field is changing rapidly and this information may all be really different in 4-5 years).
In my experience (I work in tech, but I am not a data scientist, although I do work with some) is that the exact language you know is less important than knowing one or more. The view that most tech companies take when hiring data scientists is that smart people can learn a new programming language fast. For example, a current ad for a data scientist at Microsoft asks for “software development skills in one or more high level languages (C#/C/C++/Java/F#), one or more scripting languages (Python/Perl/Shell), knowledge of R. Expertise any of the common ML tools (Weka, R, RapidMiner, KNIME, Scikit Learn, SAS, SPSS, ModelBuilder etc.) is a plus.” What they really want are smart people with the statistical chops and the creativity to solve problems in interesting ways.
However, I will say that there are a few things that most ads ask for: knowledge of R (any statistics department worth their salt is teaching this now, though); knowledge of SQL (friends in the know say that this is going to change in the next 4-5 years, but as of right now, almost every data scientist ad I see asks for this), and knowledge of Tableau as a data visualization tool. Many ask for expertise in Excel, and some others want some SAS knowledge. Basically, the more exposure you can get to programming languages and statistical analysis tools the better, and there are probably ones that come up more often (C++, Python, Java) but focus less on specifics and more on acquiring some skills and the ability to learn rapidly and teach yourself.
Just study what you want/like to study and maybe revisit your grad school chances in a year or two. It’s best to wait until you’ve had more of a taste of your field and until you’ve gotten to know your professors and can get the low-down on your field and your admissions prospects.