- the job market for the language
- the long term prospects for the language
- how easy the language is to learn
- what projects you can build while you’re learning (and share with friends so you can stay motivated)
...and he's right.
If you're interested in Data Science and want to learn a language, look at your job market. See what kind of of programming experience they're looking for. Is it Python? Java? R? RapidMiner? Then select two and become master at one.
You're going to have to spend 1000's of hours learning one particular language (or tool) to become 'job market' proficient so choose wisely.