Even dealing with the relatively small gigabytes worth of data that I have to deal with, I spend a lot of time massaging data sets so that they can all be corralled and used effectively together.
I imagine that large corporations that deal with terabytes or petabytes of data have rooms filled with H-1B drones that write dozens of data converters so that a few math PhD's can work their wonders on the aggregated results.
Even big data doesn't sound like a likely career path for the number of STEM students graduating each year.
The problem is that H-1Bs have a short shelf life and as soon as they get their experience they move on which is the real reason why corps are pushing so hard to get even more authorized. The company I work for has actually had to move a lot of its offshore operations to the Philippines due the the increasing cost of consultants in India.
What large corporations have (and I’ve worked for the largest ones in the US) are ETL teams and BI teams for ‘massaging data’. The truth is having a bunch of cheap labor doing transformation mappings doesn’t work as well in practice as it does in theory and is pretty inefficient and numbers of cheap staff don’t always accomplish what real experienced staff can.
Right now the biggest barrier to corporations hiring in the US for what would be entry level jobs is the tax code. I’m sure many young kids out of school would be happy to take entry level tech jobs. The problem is the federal government has made it prohibitive. The biggest area where this is evident is Manufacturing. Apple could be making phones in the US rather than China if not for the 35% corporate tax and the stupid rules on taxes overseas profits. So what happens is the jobs that should be going to lower middle class and poor kids getting them on a path to a long career instead are going to communist Chinese and what do the US kids get? They get a future of subsistence.