Really ugly and rough code to sort some content out of Excel

gistfile1.py

from collections import defaultdict
import operator
import os
from xlrd import open_workbook

"""
starting from the "raw" data, which I'm assuming is the first tab, Sorted by Customer
group everything by part number and get the total sold
then list the items in descending order of popularity
Do you need any other data beyond part and total sold? Looks like I should sum the Extended Price column as well?
"""

COL_QUANTITY = 1
COL_ITEM = 2
COL_DESCRIPTION = 9
COL_EXTENDED_PRICE = 11

wb = open_workbook('ytd_june.xls')
quantities = defaultdict(lambda: [0, 0.0, ''])
by_customer = wb.sheets()[0]
for row in range(1, by_customer.nrows):
    qty = int(by_customer.cell(row, COL_QUANTITY).value)
    item = str(by_customer.cell(row, COL_ITEM).value).strip()
    price = float(by_customer.cell(row, COL_EXTENDED_PRICE).value)
    current = quantities[item]
    quantities[item][0] = current[0] + qty
    quantities[item][1] = quantities[item][1] + price
    quantities[item][2] = str(by_customer.cell(row, COL_DESCRIPTION).value).strip()

raw = []
for k in quantities:
    data = quantities[k]
    raw.append([k, data[0], data[1], data[2]])


f = open('sorted.csv', 'w')
f.write('Part,Description,Count,Total%s' % os.linesep)
for row in raw:
    f.write('%s,%s,%d,%.2f%s' % (row[0], row[3], row[1], row[2], os.linesep))