commute.py

import json
from datetime import datetime
from collections import namedtuple
from itertools import groupby, takewhile
from statistics import median, mean
from matplotlib import pyplot
import numpy as np
from geopy.distance import geodesic

Point = namedtuple('Point', 'latitude, longitude, datetime')


def read_points():
    with open('data.json') as f:
        data = json.load(f)

    for point in data['locations']:
        yield Point(
            point['latitudeE7'] / 10 ** 7,
            point['longitudeE7'] / 10 ** 7,
            datetime.fromtimestamp(int(point['timestampMs']) / 1000)
        )


home = (52.350879, 4.893817)
work = (52.3657573, 4.8980648)

max_distance = 0.050

from_date = datetime(2017, 11, 1)
from_hour = 7
to_hour = 12

points = read_points()

after_move = takewhile(lambda point: point.datetime >= from_date, points)

work_days = (point for point in after_move
             if point.datetime.weekday() < 5)

commute_time = (point for point in work_days
                if from_hour <= point.datetime.hour < to_hour)

by_days = groupby(commute_time, key=lambda point: point.datetime.date())


def last_at_home(points):
    result = None
    for point in points:
        if geodesic(home, point[:2]).km <= max_distance:
            result = point
    return result


def first_at_work(points, after):
    for point in points:
        if point.datetime > after.datetime and geodesic(work, point[:2]).km <= max_distance:
            return point


Commute = namedtuple('Commute', 'day, start, end, took')


def get_commute():
    for day, points in by_days:
        points = [*points][::-1]

        start = last_at_home(points)
        if start is None:
            continue

        end = first_at_work(points, start)
        if end is None:
            continue

        yield Commute(
            day, start.datetime, end.datetime, end.datetime - start.datetime,
        )


commutes = [*get_commute()][::-1]

fig, ax = pyplot.subplots()
ax.plot([commute.day for commute in commutes],
        [commute.took.total_seconds() / 60 for commute in commutes])

ax.set(xlabel='day', ylabel='commute (minutes)',
       title='Daily commute')
ax.grid()
pyplot.show()

# http://www.knmi.nl/nederland-nu/klimatologie/daggegevens
from dateutil.parser import parse

Weather = namedtuple('Weather', 'windspeed, temperature, wind_direction')


def read_weather():
    result = {}

    with open('weather.txt') as f:
        for line in f.readlines():
            if not line.startswith(' '):
                continue

            data = [part.strip() for part in line.split(',')]
            result[parse(data[1]).date()] = Weather(
                int(data[4]) / 10,
                int(data[11]) / 10,
                int(data[2]),
            )

    return result

weather = read_weather()
normalized = [commute for commute in commutes
              if commute.took.total_seconds() < 60 * 20]

fig, ax = pyplot.subplots()
ax.grid()
ax.scatter([commute.took.total_seconds() / 60 for commute in normalized],
           [weather[commute.day].windspeed for commute in normalized])
ax.set(xlabel='Commute time', ylabel='Wind speed',
       title='Commute and wind')
ax.legend()
pyplot.show()

from mpl_toolkits.mplot3d import Axes3D

fig, ax = pyplot.subplots(subplot_kw={'projection': '3d'})
ax.grid()
ax.scatter([weather[commute.day].temperature for commute in normalized],
           [weather[commute.day].windspeed for commute in normalized],
           [commute.took.total_seconds() / 60 for commute in normalized])
ax.set(xlabel='Wind speed', ylabel='Temperature', zlabel='Commute time',
       title='Commute and weather')
ax.legend()
pyplot.show()

from matplotlib import cm

colors = iter(cm.Reds(np.linspace(0, 1, len(normalized))))

fig, ax = pyplot.subplots()
ax.grid()

for commute in sorted(normalized, key=lambda commute: commute.took.total_seconds() / 60):
    ax.scatter(weather[commute.day].windspeed,
               weather[commute.day].wind_direction,
               color=next(colors))

ax.set(xlabel='Wind speed', ylabel='Wind direction',
       title='Commute and wind')

ax.grid()

pyplot.show()


from matplotlib import cm

colors = iter(cm.Reds(np.linspace(0, 1, len(normalized))))

fig, ax = pyplot.subplots(subplot_kw={'projection': 'polar'})

for commute in sorted(normalized, key=lambda commute: commute.took.total_seconds() / 60):
    ax.scatter(weather[commute.day].wind_direction,
               weather[commute.day].windspeed,
               color=next(colors))

ax.set(title='Commute and wind')

ax.grid()

pyplot.show()