Interesting to note that BlackRock's Aladdin software platform, powering much of this stuff, is being largely written / re-written in Julia.
As for equivalents to Pandas, Matplotlib etc., Julia has thin wrappers around the Python versions of these so you can use them just like in Python with the same performance.
Meaning?
Just take any function you write and write code_native( your_function, () )
Knowing that the JIT is actually executing means a lot to me. For investment banks this means lower cost (quants can write code that goes into production) and lower latencies.