Looping through URLs to create csv and dataframe for pandas

Question

I want to pull data from URLs that meet specific date requirements that are shown in the URL structure, and put that information into a csvs to be used locally.

http://web.mta.info/developers/data/nyct/turnstile/turnstile_190629.txt

The series of 6 digits at the end of the URL are year-month-day indicators: 190629

I am collecting the data from March through June (03-06) for years 2016 - 2019 (16-19). If the URL exists, create a csv and also combine them all into a single csv to feed into a pandas dataframe.

This works, but it's suuuuuper slow, and I know that it's not the most pythonic way of doing this.

import requests
import pandas as pd
import itertools

date_list = [['16', '17', '18', '19'],['03', '04', '05', '06'],['01', '02', '03', '03', '04', '05', '06'
       ,'07', '08', '09','10', '11', '12','13','14' ,'15', '16',
       '17','18','19','20','21','22','23','24','25','26','27'
       ,'28','29','30','31']]
date_combo = []

# - create year - month - date combos
# - link: https://stackoverflow.com/questions/798854/all-combinations-of-a-list-of-lists

for sub_list in itertools.product(*date_list):
    date_combo.append(sub_list)

url_lead = 'http://web.mta.info/developers/data/nyct/turnstile/turnstile_'
url_list = []

# - this checks the url is valid and adds them to a list
for year, month, day in date_combo:
    concat_url = url_lead + year + month + day + '.txt'
    response = requests.get(concat_url)
    if response.status_code == 200:
# ---- creates a list of active urls
        url_list.append(concat_url)
# ---- this creates individual csvs ---- change path for saving locally
# ---- filename is date
        df = pd.read_csv(concat_url, header = 0, sep = ',')
        df.to_csv(r'/Users/.../GitHub/' + year + month + day + '.csv')

# - this creates a master df for all urls
dfs = [pd.read_csv(url,header = 0, sep = ',') for url in url_list]
df = pd.concat(dfs, ignore_index = True)
df.to_csv(r'/Users/.../GitHub/seasonal_mta_data_01.csv')

My code is running as expected, but I'd appreciate any recommendations to clean it up!

Quang Hoang · Accepted Answer · 2019-07-04T19:23:21.453

Not much that I can think of. Here are some things I would do differently:

# more consie construction of date_combo  
date_list = [range(16,20), range(3,7),range(1,32)]
date_combo = [sub_list for sub_list in itertools.product(*date_list)]

url_lead = 'http://web.mta.info/developers/data/nyct/turnstile/turnstile_'
url_list = []
dfs = []

# - this checks the url is valid and adds them to a list
for year, month, day in date_combo:
    # year, month, day are integers
    # so we use f string here
    concat_url = f'{url_lead}{year}{month:02}{day:02}.txt'

    response = requests.get(concat_url)
    if response.status_code == 200:
        url_list.append(concat_url)

        # append to dfs and save csv
        dfs.append(pd.read_csv(concat_url, header = 0, sep = ','))
        dfs[-1].to_csv(f'/Users/.../GitHub/{year}{month:02}{day:02}.csv)

# we don't need to request the txt files again:
df = pd.concat(dfs, ignore_index = True)
df.to_csv(r'/Users/.../GitHub/seasonal_mta_data_01.csv')

I got a TypeError from concatenating the ints: TypeError: can only concatenate str (not "int") to str — KMB, Jul 04 '19 at 19:21
My bad, you would need to do similarly to `concat_url`: `f'/Users/.../GitHub/{year}{month:02}{day:02}.csv'` — Quang Hoang, Jul 04 '19 at 19:22

Looping through URLs to create csv and dataframe for pandas

1 Answers1