New instance of a class contains older data given to an older instance

I'm trying to analyze some reviews using vader in python. I made an Analyzer class like so:

class Analyzer:
    dataframe = None
    reviews = []
    
    # other data

    def __init__(self, brand):
        self.dataframe = pd.read_csv(brand) # using pandas
        reviews = self.dataframe['Reviews'].tolist()

    def analyze(self):
        # code for analyze...

    # more codes...

and I used this class in main.py like so:

from analyzer import Analyzer

brands = [
    'apple',
    'google',
    'huawei',
    'motorola',
    'nokia',
    'samsung',
    'sony',
    'xiaomi'
]

for brand in brands:
    analysis = Analyzer(brand)
    analysis.analyze()
    del analysis

now the problem is: When brands are given to the class to be analyzed, the older reviews remain in the list.

e.g.: apple.csv has 1000 reviews and google.csv has 700 reviews. but when google is passed to analyzer, the reviews list length is not 700, but 1700.

reviews attribute is for class and not object, since you defined it in the body of the class.
define the reviews in the body of __init__:

class Analyzer:
    
    # other data

    def __init__(self, brand):
        self.dataframe = pd.read_csv(brand) # using pandas
        self.reviews = self.dataframe['Reviews'].tolist()

    def analyze(self):
        # code for analyze...

    # more codes...