I use it often. Even after 20 years of using numeric/numarray/numpy, I still feel more productive in Octave and Matlab. Mostly time series analysis, signal processing, image analysis. Occasional modeling and simulation work. I think for me it’s that octave does it’s job well and doesn’t try to do other things. Python grates on me occasionally since it’s libraries seem to lack focus and try to be too much for too broad of an audience. I like focused, simple tools. I do use Python, but not as my sole tool. I don’t understand he mindset that one tool/language should do everything. Easier for me to learn many languages than one language and have to learn many independently designed libraries.