Used for creating histogram of thimble.io user signups

plot_signups_histogram.py

#!/usr/bin/env python2

import csv
from dateutil import tz
import dateutil.parser
import numpy as np
import pylab as P
from matplotlib.dates import MinuteLocator, HourLocator, DayLocator, DateFormatter
from matplotlib.dates import date2num, num2date

timestamps = []
with open('Thimble.csv', 'rb') as csvfile:
    # Thimble.csv field names: id,form_id,created_at,spam,email,name
    csvreader = csv.reader(csvfile, delimiter=',')
    # skip header row
    csvreader.next()
    for row in csvreader:
        # signup_time = created_at field
        signup_time = dateutil.parser.parse(row[2])
        from_zone = tz.gettz('UTC')
        to_zone = tz.gettz('America/New_York')
        utc = signup_time.replace(tzinfo=from_zone)
        # strip all timezone data so we're showing NY time
        timestamps.append(utc.astimezone(to_zone).replace(tzinfo=None))

(hist, bin_edges) = np.histogram(date2num(timestamps), bins=48, density=False, normed=False)
width = bin_edges[1] - bin_edges[0]
fig = P.figure()
ax = fig.add_subplot(111)
ax.bar(bin_edges[:-1], hist, width=width)
ax.set_ylabel('New Subscribers')
ax.set_xlabel('MM/DD/YY - HH:MM (New York timezone)')
ax.set_title('New Subscribers per Hour')
ax.xaxis.set_major_locator(HourLocator(interval=2))
ax.xaxis.set_major_formatter(DateFormatter('%D - %H:%M'))
fig.autofmt_xdate()

P.show()