parking and trip charting/plotting code 2016-08-25

analyze.py

#!/usr/bin/env python2
# coding=utf-8

from __future__ import print_function
import sys
import simplejson as json
import datetime

import numpy as np
import matplotlib.pyplot as plt


sys.path.insert(0, '/home/jarek/projects/electric2go')

from electric2go.systems.car2go import city
from electric2go.analysis import graph


def json_deserializer(obj):
    # parse datetimes from JSON we wrote
    for (key, value) in obj.items():
        if isinstance(value, basestring) and key.endswith('time'):
            try:
                # this is the format that isoformat outputs
                obj[key] = datetime.datetime.strptime(value, "%Y-%m-%dT%H:%M:%S")
            except:
                pass

    return obj


def adjust_tz(t, tz):
    return (t - tz) if t > tz else (t + 24 - tz)


def get_time(t, tz):
    hours_plus_minutes = t.hour + (t.minute/60.0)
    return adjust_tz(hours_plus_minutes, tz)


result_dict = json.load(fp=sys.stdin, object_hook=json_deserializer)


system = result_dict['metadata']['system']


tz = 2  # Berlin in summertime


all_parkings = [parking for vin in result_dict['finished_parkings'] for parking in result_dict['finished_parkings'][vin]
                if parking['starting_time'] != result_dict['metadata']['starting_time']]
                # don't include parkings that are detected to have started on the very first data point - likely that they've actually started earlier
                # and it would be a misleading data point
all_trips = [trip for vin in result_dict['finished_trips'] for trip in result_dict['finished_trips'][vin]]


def parking_cleaner(parking):
    return True
    # or, for instance, use `parking['duration'] < 60*60*24` to only show things lasting less than a day - use when we want to eliminate outliers


all_parking_durations = [p['duration']/60 for p in all_parkings if parking_cleaner(p)]
all_parking_starting_hour = [get_time(p['starting_time'], tz) for p in all_parkings if parking_cleaner(p)]
all_parking_ending_hour = [get_time(p['ending_time'], tz) for p in all_parkings if parking_cleaner(p)]
all_parking_fuel = [p['fuel'] for p in all_parkings if parking_cleaner(p)]



fig = plt.figure()                                                     
ax = fig.add_subplot(3,1,1)  
plt.hist([p for p in all_parking_durations if p < 400], bins=400, histtype='bar')
plt.title(system + " parking period duration (minutes, < 400 min)")
plt.ylabel("frequency")

ax = fig.add_subplot(3,1,2)
plt.hist(all_parking_ending_hour, bins=288, histtype='bar')
plt.title(system + " parking ending time of day")
plt.ylabel("frequency")

ax = fig.add_subplot(3,1,3)
plt.hist(all_parking_starting_hour, bins=288, histtype='bar')
plt.title(system + " parking starting time of day")
plt.ylabel("frequency")

plt.show()

fig = plt.figure()
plt.plot(all_parking_fuel, all_parking_durations, '.')
plt.title(system + " parking period duration vs reported fuel level")
plt.xlabel("reported fuel level (%)")
plt.ylabel("parking duration (minutes)")
plt.show()


fig = plt.figure()
plt.plot(all_parking_starting_hour, all_parking_durations, '.')
plt.title(system + " parking period duration vs starting time of the parking period")
plt.xlabel("time the parking period started (time of day)")
plt.ylabel("parking duration (minutes)")
plt.show()


fig = plt.figure()
plt.plot(all_parking_ending_hour, all_parking_durations, '.')
plt.title(system + " parking period duration vs ending time of the parking period")
plt.xlabel("time the parking period ended (time of day)")
plt.ylabel("parking duration (minutes)")
plt.show()



max_minutes = 60*6

def filterer(trip):
    return trip['duration'] < 60*max_minutes #and trip['starting_time'].hour == 10


all_trip_lengths = [trip['duration']/60 for trip in all_trips if filterer(trip) ]
all_trip_speeds = [trip['speed'] for trip in all_trips if filterer(trip) and trip['speed'] < 50 ]
all_trip_distances = [trip['distance'] for trip in all_trips if filterer(trip) ]
all_trip_fuels = [trip['fuel_use'] for trip in all_trips if filterer(trip) ]#and trip['fuel_use'] > -20 ]


fig = plt.figure()
plt.plot(all_trip_lengths, all_trip_distances, '.')
plt.title(system + " trip duration vs distance")
plt.xlabel("duration (minutes)")
plt.ylabel("distance")
# also works as:
#all_trip_duration_vs_distance = [(trip['duration']/60, trip['distance']) for trip in all_trips if trip['duration'] < 60*max_minutes ]
#dur,dist = zip(*all_trip_duration_vs_distance)
plt.show()


fig = plt.figure()
plt.plot(all_trip_lengths, all_trip_fuels, '.')
plt.title(system + " trip duration vs fuel use")
plt.xlabel("duration (minutes)")
plt.ylabel("fuel use")
plt.show()




fig = plt.figure()                                                     
ax = fig.add_subplot(4,1,1)  
plt.hist(all_trip_lengths, bins=max_minutes-1, histtype='bar')
plt.title(system + " trip durations (minutes)")
plt.ylabel("frequency")

ax = fig.add_subplot(4,1,2)
plt.hist(all_trip_distances, bins=max_minutes-1, histtype='bar')
plt.title(system + " trip distances (km)")
plt.ylabel("frequency")

ax = fig.add_subplot(4,1,3)
plt.hist(all_trip_speeds, bins=max_minutes-1, histtype='bar')
plt.title(system + " trip speeds (km/h)")
plt.ylabel("frequency")

ax = fig.add_subplot(4,1,4)
plt.hist(all_trip_fuels, bins=max_minutes-1, histtype='bar')
plt.title(system + " fuel usage (%)")
plt.ylabel("frequency")

plt.show()


def non_zero_distance_filterer(trip):
	return filterer(trip) and trip['distance'] > 0


filtered_trip_lengths = [trip['duration']/60 for trip in all_trips if non_zero_distance_filterer(trip) ]
filtered_trip_speeds = [trip['speed'] for trip in all_trips if non_zero_distance_filterer(trip) and trip['speed'] < 50 ]
filtered_trip_distances = [trip['distance'] for trip in all_trips if non_zero_distance_filterer(trip) ]
filtered_trip_fuels = [trip['fuel_use'] for trip in all_trips if non_zero_distance_filterer(trip) and trip['fuel_use'] > -20 ]

fig = plt.figure()                                                     
ax = fig.add_subplot(4,1,1)  
plt.hist(filtered_trip_lengths, bins=max_minutes-1, histtype='bar')
plt.title(system + " non-zero distance trip durations (minutes)")
plt.ylabel("frequency")

ax = fig.add_subplot(4,1,2)
plt.hist(filtered_trip_speeds, bins=max_minutes-1, histtype='bar')
plt.title(system + " non-zero trip distances (km)")
plt.ylabel("frequency")

ax = fig.add_subplot(4,1,3)
plt.hist(filtered_trip_distances, bins=max_minutes-1, histtype='bar')
plt.title(system + " non-zero trip speeds (km/h)")
plt.ylabel("frequency")

ax = fig.add_subplot(4,1,4)
plt.hist(filtered_trip_fuels, bins=max_minutes-1, histtype='bar')
plt.title(system + " non-zero distance fuel usage (%)")
plt.ylabel("frequency")

plt.show()