root_numpy is a Python extension for converting ROOT TTrees into NumPy recarrays or structured arrays. With the core internals written in C++, root_numpy can efficiently handle large amounts of data (limited only by the available memory). Now that your ROOT data is in NumPy form, you can make use of the many powerful scientific Python packages or perform quick exploratory data analysis in interactive environments like IPython (especially IPython's popular notebook feature).
root_numpy currently supports basic types like Bool_t, Int_t, Float_t, Double_t, etc. and arrays of basic types (both variable and fixed-length). Vectors of basic types are also supported.
Tab completion for numpy.recarray column names (yourdata.<TAB>) is also available with this numpy extension.
The rootpy package also provides a script that uses root_numpy and PyTables to convert all TTrees in a ROOT file into the HDF5 format.
root_numpy is tested with ROOT 5.32, NumPy 1.6.1, Python 2.7.1 but it should work in most places.
python setup.py install