freq_distributions.py

import pandas

variables_of_interest = {
    'S3BQ1A5': ('ever used cannabis', {1: 'Yes', 2: 'No', 9: 'Unknown'}),
    'S3BQ1A6': ('ever used cocaine/crack', {1: 'Yes', 2: 'No', 9: 'Unknown'}),
    'S3BQ1A9A': ('ever used heroin', {1: 'Yes', 2: 'No', 9: 'Unknown'}),
    'S3CD5Q13A': ('age at onset of cannabis abuse', {
        '5-64': 'Age',
        99: 'Unknown',
        'BL': 'NA, didnt meet symptom criteria for lifetime cannabis abuse'}),
    'S3CD6Q13A': ('age at onset of cocaine abuse', {
        '5-52': 'Age',
        99: 'Unknown',
        'BL': 'NA, didnt meet symptom criteria for lifetime cocaine abuse'}),
    'S3CD9Q13A': ('age at onset of heroin abuse', {
        '9-62': 'Age',
        99: 'Unknown',
        'BL': 'NA, didnt meet symptom criteria for lifetime heroin abuse'}),
    'S3BD5Q2F': ('age when began using cannabis most', {
        '5-64': 'Age',
        99: 'Unknown',
        'BL': 'NA, never or unknown if ever used cannabis'}),
    'S3BD6Q2F': ('age when began using cocaine or crack most', {
        '5-64': 'Age',
        99: 'Unknown',
        'BL': 'NA, never or unknown if ever used cocaine or crack'}),
    'S3BD9Q2F': ('age when began using heroin most', {
        '12-62': 'Age',
        99: 'Unknown',
        'BL': 'NA, never or unknown if ever used heroin'})
}


def load_data():
    pandas.set_option('display.float_format', lambda x: '%f' % x)

    data = pandas.read_csv('nesarc_pds.csv', low_memory=False)
    data.columns = map(str.upper, data.columns)

    print ('# of observations is ' + str(len(data)))
    print ('# of variables is ' + str(len(data.columns)))

    return data


def concat_description(tuple_value):
    result = tuple_value[0] + '( '
    for key, value in tuple_value[1].iteritems():
        result += str(key) + ' : ' + value + ' '
    result += ')'
    return result


def main():
    data = load_data()

    for code in variables_of_interest:
        data[code] = data[code].convert_objects(convert_numeric=True)

    for code, tuple_value in variables_of_interest.iteritems():
        description = concat_description(tuple_value)
        print ('counts for ' + code + ' -- ' + description)
        counts = data[code].value_counts(sort=False)
        print (counts)

        print ('percentages for ' + code + ' -- ' + description)
        percentages = data[code].value_counts(sort=False, normalize=True)
        print (percentages)

if __name__ == 'main':
    main()