Strange behaviour of Python list used as class attribute

Question

I have created a Python class like the following:

class Node(object):
    """docstring for Node"""
    def __init__(self, node_id, lng, lat, gid, raw_data=[],quantized_data=[]):
        super(Node, self).__init__()
        self.node_id = node_id
        self.lng = lng
        self.lat = lat
        self.gid = gid
        self.raw_data = raw_data
        self.quantized_data = quantized_data

Class Node has two attributes with default values, which are empty lists. Then I create two objects:

a = Node(1,1,1,1)
b = Node(2,2,2,2)

I find out that if I change raw_data of one object, the other also changes. For example:

a.raw_data.append('x')

Then b.raw_data also has x:

>>> b.__dict__
{'node_id': 2, 'lng': 2, 'lat': 2, 'gid': 2, 'raw_data': ['x'], 'quantized_data': []}

I wonder why changing the attribute of one object affects the other. Obviously, they are two separate objects. What's wrong here or there's something I don't understand? Thanks