Here's an example, along with a few support libraries, on how to work with data from the synchronized brainwave dataset.

Download the dataset, name the folder "dataset" and place it in the root directory of this project to run the example.

feature_vector_generator(subject_num, time0, time1)

Returns a generator of feature vectors for the given subject between time0 and time1.

Optionally, this can take a third argument, sq, which defines a threshold signal quality to be eligible in a feature vector. (By default, only readings with perfect signal quality are included).

Of particular note is lib/featurevectorgenerator.py, which contains most of logic around building feature vectors. (The rest is from @wazaahhh's brainlib).

Running example.py requires scikit-learn and numpy, plus dateutil for parsing dates from the source .csv files. You should be able to pip install all of these on your platform.