-
Notifications
You must be signed in to change notification settings - Fork 22
/
plotter_new.py
1265 lines (1043 loc) · 44.4 KB
/
plotter_new.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
from utils import *
import report_card
import matplotlib
#matplotlib.use('Agg')
from matplotlib.backends.backend_pdf import PdfPages
import ephem
import copy
import math
import sys
import numpy as np
from datetime import datetime
import time as itime
import matplotlib.pyplot as plt
from matplotlib.colors import LogNorm
import cPickle as pickle
import pytz
import heapq
from sklearn.cluster import KMeans
from sklearn import mixture
from holiday import yfhol
import warnings
utc_tz = pytz.utc
tz_used = pytz.timezone("US/Central")
#tz_used = pytz.timezone("America/Chicago")
def getSun(stateID, currentTime, city = None):
"""Get the position of the sun at a given time and location.
Parameters:
stateID -- A string abbreviation of the state name, ie "IL","AZ",etc..
currentTime -- Local time with tzinfo = state time zone.
city (optional) -- The City (defaults to state capital if missing or not found in database).
Returns:
sin(altitude) (representing how much sunlight hits an area)
"""
#Note: altitude,azimuth are given in radians
o = ephem.Observer()
stateID.capitalize()
dateStamp=currentTime.astimezone(pytz.utc)
if city is None:
city=states[stateID]['capital']
else:
try:
city=states[stateID][city].capitalize()
except:
city=states[stateID]['capital']
the_lat = states[stateID][city][0]
the_long = states[stateID][city][1]
#IF YOU CONVERT TO FLOAT HERE, EVERYTHING GOES BOOM (but quietly)
with warnings.catch_warnings():
warnings.simplefilter("ignore")
o.lat = the_lat
o.long = the_long
o.date = dateStamp
sun = ephem.Sun(o)
alt=sun.alt
return math.sin(alt)
def get_periods(di, nobs, first_pred, which = "kwhs", skip_fun = (lambda x: False), wrap_around = False):
"""Get a collection of periods (e.g., weeks) from a building record.
Parameters:
di -- The building record.
nobs -- The number (int) of observations for a single period (e.g., 168 for weeks).
first_pred -- A function which, given a datetime object, returns True if it is the start of the period.
For example, first_pred returns True only when the argument is Sunday at Midnight.
which -- A string representing which time series in the building record to use (defaults to "kwhs")
skip_fun -- A function which takes a datetime object and returns True if that time should be skipped.
For example, skip_fun may return True on weekends, to obtain only work weeks.
wrap_around -- If True, the beginning part of the time series is placed at the end.
For example, if we're starting periods on Monday, but the first day in the time series
is Thursday, then the first part (from Thrusday to Monday) will be moved to the end.
Returns (pers, new_times):
pers -- The values (e.g., kwhs) of the periods.
new_times -- The times (datetime objects) associated with the values.
"""
d = copy.deepcopy(di) #used so that we don't change di at all
times = d["times"]
series, series_oriflag = d[which]
additional_mask = np.array([skip_fun(t) for t in times])
new_mask = np.logical_or((~series_oriflag), additional_mask)
masked_series = np.ma.array(series, mask = new_mask)
first = 0
for ind, t in enumerate(times):
if first_pred(t):
first = ind
break
if wrap_around:
pers = np.ma.concatenate([masked_series[first:], masked_series[:first]])
new_times = np.ma.concatenate([times[first:], times[:first]])
else:
pers = masked_series[first:]
new_times = times[first:]
residue = len(pers) % nobs
if residue != 0:
pers = pers[:-residue]#trim off extra
new_times = new_times[:-residue]
pers = pers.reshape(-1, nobs)
new_times = new_times.reshape(-1, nobs)
return pers, new_times
def make_text_fig(d, textfig):
"""Show the static information from the building record in a given axis.
Parameters:
d -- The building record.
textfig -- The axis to hold the figure.
"""
bid = d["bid"]
naics = d["naics"]
btype = d["btype"]
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
temps, temps_oriflag = d["temps"]
naics_map, desc = qload("NAICS.pkl")
if naics in naics_map:
naics_str = naics_map[naics]
naics_str = " (" + naics_str.lstrip().rstrip() + ")"
else:
naics_str = ""
if len(naics_str) > 20:
naics_str = naics_str.replace("and ", "and\n") #TODO: Figure out better way
toPrint = "ID:\n " + str(bid)\
+ "\nNaics:\n " + str(naics) + "\n " + naics_str\
+ "\nType:\n " + str(btype)
toPrint += "\n"
rep = report_card.get_report(d)
width = 50
for k in rep.keys():
toPrint += k + ":\n" + ("%.2f" % rep[k]).rjust(width) + "\n"
textfig.text(0.05, .95, toPrint, fontsize=10, ha='left', va='top')
textfig.set_xticks([])
textfig.set_yticks([])
def make_prison_text_fig(d, textfig):
"""Show the static information from the building record in a given axis.
Note: This function relies on the building record corresponding to a prison in our database.
Parameters:
d -- The building record.
textfig -- The axis to hold the figure.
"""
bid = d["bid"]
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
temps, temps_oriflag = d["temps"]
feature_map, desc = qload("prison_features_map.pkl")
if isinstance(bid, str) and "q" in bid:
features = feature_map[int(bid[:-2])]
else:
features = feature_map[bid]
toPrint = features["name"] + " \n (" + features["acronym"] + ")"\
+ "\nCity:\n " + features["city"]\
+ "\nID:\n " + str(bid)\
+ "\nNumber of Buildings:\n " + features["nrBuildings"]\
+ "\nX-houses:\n " + features["X_design_housing_units"]\
+ "\nT-houses:\n " + features["T_type_housing_units"]\
+ "\nHealthcare units:\n " + features["health_care_units"]\
+ "\nBuilding size:\n " + features["building_size"]
for i in range(1, 7):
iname = "industry_%d"%i
industry = features[iname]
if len(industry) > 15:
industry = industry.replace(" and", "\nand")
if industry != "":
toPrint += "\nIndustry:\n " + industry
toPrint +="\nTotal land:\n " + features["total_land"] + "acres"\
+ "\nAverage Hourly Energy Usage:\n " + str(np.round(np.average(kwhs), 2)) + "kw"\
+ "\nMin:\n " + str(np.round(np.min(kwhs), 2))\
+ "\nMax:\n " + str(np.round(np.max(kwhs), 2))\
+ "\nTotal:\n " + str(np.round(np.sum(kwhs)/1000.0, 2)) + "gwh"
toPrint += "\n"
width = 50
textfig.text(0.05, .95, toPrint, fontsize=10, ha='left', va='top')
textfig.set_xticks([])
textfig.set_yticks([])
def make_temp_vs_time_fig(d, tvt):
"""Show temperature as a function of time in a given axis.
Parameters:
d -- The building record.
tvt -- The axis to hold the figure.
"""
times = d["times"]
temps, temps_oriflag = d["temps"]
tvt.plot( times[temps_oriflag] , temps[temps_oriflag] , c = "blue")
tvt.scatter(times[~temps_oriflag], temps[~temps_oriflag], edgecolors = "none", c = "red")
tvt.scatter(times[~temps_oriflag], [0 for x in temps[~temps_oriflag]], edgecolors = "none", c = "red", s = 1)
tvt.set_title("Temperature Over Time")
tvt.set_ylabel("Temperature")
labels = tvt.get_xticklabels()
for label in labels:
label.set_rotation(30)
def make_kwhs_vs_time_fig(d, tvk):
"""Show kwhs as a function of time in a given axis.
Parameters:
d -- The building record.
tvk -- The axis to hold the figure.
"""
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
tvk.plot(times[kwhs_oriflag], kwhs[kwhs_oriflag], c = "blue", label = "Energy Usage")
tvk.scatter(times[~kwhs_oriflag], kwhs[~kwhs_oriflag], c = "red", edgecolors = "none", label = "Imputed Values")
tvk.scatter(times[~kwhs_oriflag], [0 for x in kwhs[~kwhs_oriflag]], c = "red", edgecolors = "none", s = 1)
ori_kwhs = kwhs[kwhs_oriflag]
per_95_kwhs = np.percentile(ori_kwhs, 95)
per_5_kwhs = np.percentile(ori_kwhs, 5)
tvk.axhline(y = per_95_kwhs, c = "red", ls = "dashed", label = "95th Percentile")
tvk.axhline(y = per_5_kwhs, c = "black", ls = "dashed", label = "5th Percentile")
tvk.set_title("Energy Usage Over Time")
tvk.legend()
tvk.set_ylabel("kwhs")
labels = tvk.get_xticklabels()
for label in labels:
label.set_rotation(30)
def make_freqs_fig(d, freqs):
"""Show kwhs in the frequency domain (via the DFT).
Note: This function requires data from (all of) the_year.
Parameters:
d -- The building record.
freqs -- The axis to hold the figure.
"""
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
temps, temps_oriflag = d["temps"]
start_time = "01/01/%d 00:00:00" %the_year
end_time = "01/01/%d 00:00:00" %(the_year + 1)
start_ts = int(itime.mktime(datetime.strptime(start_time, "%m/%d/%Y %H:%M:%S").timetuple()))
end_ts = int(itime.mktime(datetime.strptime(end_time, "%m/%d/%Y %H:%M:%S").timetuple()))
all_times = range(start_ts, end_ts, 3600)
num_times = len(all_times)
a = np.fft.fft(kwhs)
half = int(num_times + 1)/2
a = a[0:half +1]
a[0] = 0 #drop constant part of signal
reals = [x.real for x in a]
imags = [x.imag for x in a]
freqs.set_xlabel("Period (h)")
freqs.set_ylabel("Magnitude")
freqs.plot(reals, label = "Real", alpha = 0.9)
freqs.plot(imags, label = "Imaginary", alpha = 0.8)
highlighted_periods = np.array([3, 6, 12, 24, 168])
highlighted_freqs = float(num_times) / highlighted_periods
highlighted_labels = [str(x) for x in highlighted_periods]
freqs.set_xticks(highlighted_freqs)
freqs.set_xticklabels(highlighted_labels)
freqs.legend()
freqs.set_title("Energy usage in the frequency domain")
labels = freqs.get_xticklabels()
for label in labels:
label.set_rotation(30)
def make_temp_vs_kwh_fig(d, tmpsvk, agg_to_day = False):
"""Show a 2d-histogram of temperatures vs kwhs in a given axis.
Parameters:
d -- The building record.
tmpsvk -- The axis to hold the figure.
"""
kwhs, kwhs_oriflag = d["kwhs"]
temps, temps_oriflag = d["temps"]
both_ori = np.logical_and(temps_oriflag, kwhs_oriflag)
if agg_to_day:
#is_sunday_start = (lambda x: x.weekday() == 6 and x.hour == 0)
is_midnight = (lambda x: x.hour == 0)
days, new_times = get_periods(d, 24, is_midnight)
day_avgs = np.ma.average(days, axis = 1)
temps, new_times = get_periods(d, 24, is_midnight, which = "temps")
temp_avgs = np.ma.average(temps, axis = 1)
#tmpsvk.hist2d(temp_avgs, day_totals, bins = 50, norm = LogNorm())
tmpsvk.scatter(temp_avgs, day_avgs, alpha = .4)
tmpsvk.set_title("Energy Usage vs Temperature (agg days)")
else:
tmpsvk.hist2d(temps[both_ori], kwhs[both_ori], bins = 50, norm = LogNorm())
tmpsvk.set_title("Energy Usage vs Temperature")
#tmpsvk.set_ylim(600, 1400)
tmpsvk.set_xlabel("Temperature")
tmpsvk.set_ylabel("kwh")
tmpsvk.grid(True)
def make_avg_day_fig(d, avgday):
"""Show the average day in a given axis.
Note: imputed values are ignored.
Parameters:
d -- The building record.
avgday -- The axis to hold the figure.
"""
is_midnight = (lambda x: x.hour == 0)
days, new_times = get_periods(d, 24, is_midnight, "kwhs")
skip_weekdays = (lambda x: x.weekday() < 5)
weekends, _ = get_periods(d, 24, is_midnight, "kwhs", skip_weekdays)
skip_weekend = (lambda x: x.weekday() >= 5)
weekdays, _ = get_periods(d, 24, is_midnight, "kwhs", skip_weekend)
avg_weekend = np.ma.average(weekends, axis = 0)
avg_weekday = np.ma.average(weekdays, axis = 0)
avg_day = np.ma.average(days, axis = 0)
std_weekend = np.ma.std(weekends, axis = 0)
std_weekday = np.ma.std(weekdays, axis = 0)
std_day = np.ma.std(days, axis = 0)
avgday.errorbar([0.25 + x for x in np.arange(24)][:-1], avg_day[:-1], yerr =std_day[:-1], label = "Day", alpha = .3, fmt = None, ecolor = "blue")
avgday.errorbar([-0.25 + x for x in np.arange(24)][1:], avg_weekend[1:], yerr =std_weekend[1:], label = "Weekend", alpha = .3, fmt = None, ecolor = "red")
avgday.errorbar(np.arange(24), avg_weekday, yerr =std_weekday, label = "Weekday", alpha = .3, fmt = None, ecolor = "green")
avgday.plot(np.arange(24), avg_day, label = "Day", c = "blue")
avgday.plot(np.arange(24), avg_weekend, label = "Weekend", c = "red")
avgday.plot(np.arange(24), avg_weekday, label = "Weekday", c = "green")
avgday.set_title("Average Day")
avgday.set_ylabel("Energy Usage (kwh)")
avgday.set_xlim(-0.5, 24.5)
avgday.set_xticks(range(0, 24, 4))
avgday.grid(True)
avgday.legend()
def make_avg_week_fig(d, avgweek):
"""Show the average week in a given axis.
Note: imputed values are ignored.
Parameters:
d -- The building record.
avgweek -- The axis to hold the figure.
"""
is_sunday_start = (lambda x: x.weekday() == 6 and x.hour == 0)
weeks, new_times = get_periods(d, 168, is_sunday_start, "kwhs")
avg_week = np.ma.average(weeks, axis = 0)
std_week = np.ma.std(weeks, axis = 0)
avgweek.plot(np.arange(168), avg_week, label = "Energy Usage", c = "purple")
avgweek.errorbar(np.arange(168), avg_week, yerr = std_week, label = "Energy Usage", errorevery = 6, ecolor = "purple", fmt = None, alpha = .3)
avgweek.set_title("Average Week")
avgweek.set_ylabel("Energy Usage (kwh)")
avgweek.set_xlim(-0.5, 24.5)
avgweek.set_xticks(range(0, 169, 24))
avgweek.set_xticklabels(["Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday"])
labels = avgweek.get_xticklabels()
for label in labels:
label.set_rotation(30)
avgweek.grid(True)
avgweek.legend()
def make_hist_fig(d, hist):
"""Show a histogram of hourly energy usage.
Note: imputed values are ignored.
Parameters:
d -- The building record.
hist -- The axis to hold the figure.
"""
kwhs, kwhs_oriflag = d["kwhs"]
hist.hist(kwhs[kwhs_oriflag], bins = 50)
hist.set_title("Histogram of Energy Usage")
hist.set_xlabel("kwhs")
hist.set_ylabel("Number of hours")
def gen_peaks(d, num_peaks = 3):
"""A generatore that yields the index of the highest peaks.
Parameters:
d -- The building record.
num_peaks -- The number of peaks to be yielded.
"""
kwhs, kwhs_oriflag = d["kwhs"]
inds = np.argsort(kwhs)[-num_peaks:]
for ind in inds:
yield ind
def make_peak_fig(d, ax, ind):
"""Show the 24-hour period around the time at index ind.
Parameters:
d -- The building record.
ax -- The axis to hold the figure.
ind -- The index of the time to display.
"""
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
highest_date = times[ind]
highest_val = kwhs[ind]
leftmost = max(0, ind-12)
rightmost = min(len(kwhs), ind+12)
make_interval_plot(d, ax, leftmost, rightmost)
def make_kwh_vs_sun_fig(d, ax, agg_days = False):
"""Show a 2d-histogram of kwhs vs the position of the sun in a given axis.
Note: Imputed values are ignored.
Parameters:
d -- The building record.
ax -- The axis to hold the figure.
"""
bid = d["bid"]
naics = d["naics"]
btype = d["btype"]
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
temps, temps_oriflag = d["temps"]
sun_pos = np.array([max(-100, getSun("IL", t)) for t in times])
if agg_days:
is_midnight = (lambda x: x.hour == 0)
d["sun_pos"] = (sun_pos, np.array([True for x in sun_pos]))
days, new_times = get_periods(d, 24, is_midnight)
suns, new_times = get_periods(d, 24, is_midnight, which = "sun_pos")
day_avgs = np.ma.average(days, axis = 1)
sun_avgs = np.ma.average(suns, axis = 1)
ax.scatter(sun_avgs, day_avgs, alpha = .4)
ax.set_xlim(np.min(sun_avgs), np.max(sun_avgs))
ax.set_title("Energy Usage vs Sunlight (agg days)")
else:
ax.hist2d(sun_pos[kwhs_oriflag], kwhs[kwhs_oriflag], bins = 50, norm = LogNorm())
ax.set_title("Energy Usage vs Sunlight")
ax.set_xlim(np.min(sun_pos[kwhs_oriflag]), np.max(sun_pos[kwhs_oriflag]))
left, right = ax.get_xlim()
ax.set_xticks([left/2., 0, right/2.])
ax.set_xticklabels(["Below Horizon", "Sun rise/set", "Above Horizon"])
ax.set_xlabel("Sunlight")
ax.set_ylabel("kwh")
ax.grid(True)
def make_monthly_usage_fig(d, ax):
"""Show a barchart with the totaly electricy usage by month.
Parameters:
d -- The building record.
ax-- The axis to hold the figure.
"""
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
month_breaks = [ind for ind, t in enumerate(times) if t.day == 1 and t.hour == 0]
zeroed_kwhs = [k if flg else 0 for k, flg in zip(kwhs, kwhs_oriflag)]
month_totals = [np.sum(zeroed_kwhs[s:e]) for s, e in zip(month_breaks, month_breaks[1:] + [-1])]
ax.bar([times[i] for i in month_breaks], month_totals, width = 10)
labels = ax.get_xticklabels()
for label in labels:
label.set_rotation(30)
ax.set_title("Monthly Energy Usage")
ax.set_ylabel("kwhs")
ax.grid(True)
def make_boxplot_all_days_fig(d, ax):
times = d['times']
hr_start = 8
hr_stop = 16
kwhs, kwhs_oriflag = d["kwhs"]
in_schedule = (lambda x: hr_start <= x.hour <= hr_stop)
out_schedule = (lambda x: x.hour < hr_start or x.hour > hr_stop)
out_flag = [out_schedule(t) for t in times]
in_flag = [in_schedule(t) for t in times]
kwhs_in = kwhs[np.logical_and(kwhs_oriflag, in_flag)]
kwhs_out = kwhs[np.logical_and(kwhs_oriflag, out_flag)]
ax.set_ylabel("Energy Usage (kwh)")
labels = ['8am - 4pm', '4pm - 8am']
ax.set_xticklabels(labels)
ax.boxplot([kwhs_in, kwhs_out])
ax.yaxis.grid(True)
ax.set_title("All days")
def make_boxplot_weekday_vs_end_fig(d, ax):
times = d['times']
hr_start = 6
hr_stop = 17
kwhs, kwhs_oriflag = d["kwhs"]
in_schedule = (lambda x: hr_start <= x.hour <= hr_stop)
out_schedule = (lambda x: x.hour < hr_start or x.hour > hr_stop)
is_weekday = (lambda x: x.weekday() < 5)
is_weekend = (lambda x: x.weekday() >= 5)
out_flag = [out_schedule(t) for t in times]
in_flag = [in_schedule(t) for t in times]
weekday_flag = [is_weekday(t) for t in times]
weekend_flag = [is_weekend(t) for t in times]
weekday_kwhs_in = kwhs[np.logical_and(kwhs_oriflag, np.logical_and(in_flag, weekday_flag))]
weekday_kwhs_out = kwhs[np.logical_and(kwhs_oriflag, np.logical_and(out_flag, weekday_flag))]
weekend_kwhs_in = kwhs[np.logical_and(kwhs_oriflag, np.logical_and(in_flag, weekend_flag))]
weekend_kwhs_out = kwhs[np.logical_and(kwhs_oriflag, np.logical_and(out_flag, weekend_flag))]
ax.set_ylabel("Energy Usage (kwh)")
ax.yaxis.grid(True)
labels = ['Weekday 8am - 4pm', 'Weekend 8am - 4pm', 'Weekday 4pm - 8am', 'Weekend 4pm - 8am']
ax.set_xticklabels(labels)
ax.boxplot([weekday_kwhs_in, weekend_kwhs_in, weekday_kwhs_out, weekend_kwhs_out], positions = [1, 2, 4, 5])
ax.set_title("Weekday vs Weekend")
def gen_holidays(d):
"""A generator that yields federal holidays in the timeframe of the building record d (in chronological order)."""
times = d["times"]
holidays = yfhol(the_year)
holidays = holidays.items()
holinames, holidates = zip(*holidays)
mymap = dict(zip(holidates, holinames))
is_midnight = (lambda x: x.hour == 0)
days, new_times = get_periods(d, 24, is_midnight, "kwhs")
for t in new_times:
date = t[0].date()
if date in holidates:
left_side = np.argmax(times == t[0])
right_side = np.argmax(times == t[-1])
yield (left_side, right_side), mymap[date]
def gen_strange_pers(d, num_pers = 3, period = "day"):
"""A generator which yields the strangest period (day or week).
Parameters:
d -- The building record.
num_pers -- The number of periods to be yieled (one at a time).
period -- A string, either "day" or "week".
"""
kwhs, kwhs_oriflag = d["kwhs"]
times = d["times"]
if period == "day":
first_pred = (lambda x: x.hour == 0)
elif period == "week":
first_pred = (lambda x: x.weekday() == 0 and x.hour == 0)
else:
print "period must be 'day' or 'week'."
return
num_per_period = 24 if period == "day" else 168
pers, new_times = get_periods(d, num_per_period, first_pred, "kwhs")
temp_pers, _ = get_periods(d, num_per_period, first_pred, which = "temps")
avg_per = np.average(pers, axis=0)
weirdness = []
totals = []
standardize = True
if standardize:
avg_per = avg_per - np.min(avg_per)
avg_per = avg_per / np.max(avg_per)
for per in pers:
if standardize:
per = per - np.min(per)
per = per / np.max(per)
dist = np.average(np.abs(per - avg_per))
weirdness.append(dist)
inds = np.argsort(weirdness)[-num_pers:][::-1]
for ind in inds:
left_side = np.argmax(times == new_times[ind][0])#hackish, but oh well
right_side = np.argmax(times == new_times[ind][-1])#hackish, but oh well
yield left_side, right_side
def make_strange_per_fig(d, ax, per, c = 'blue'):
"""Creates a plot of the period yieled from get_strange_pers.
Parameters:
d -- The building record.
ax -- The axis to hold the figure.
per -- The period yieled from get_strange_pers.
"""
start, end = per
make_interval_plot(d, ax, start, end, c = c)
ax.set_title(d["times"][start].strftime("%m/%d/%Y %H:%M:%S") + "--" + d["times"][end].strftime("%m/%d/%Y %H:%M:%S"))
def make_extreme_days_figs(d, axhigh, axlow):
"""Show the extreme high and extreme low days (in terms of electricity usage).
Parameters:
d -- The building recorod.
axhigh -- The axis to hold the extreme-high figure.
axlow -- The axis to hold the extreme-low figure.
"""
is_midnight = (lambda x: x.hour == 0)
days, new_times = get_periods(d, 24, is_midnight, "kwhs")
avg_day = np.average(days, axis=0)
weirdness = []
totals = []
times = d["times"]
for day in days:
total = np.sum(day)
totals.append(total)
dist = np.average(np.abs(day - avg_day))
weirdness.append(dist)
ind = np.argmax(totals)
highest_day = new_times[ind][0]
left_side = np.argmax(times == new_times[ind][0])#hackish, but oh well
right_side = np.argmax(times == new_times[ind][-1])#hackish, but oh well
make_interval_plot(d, axhigh, left_side, right_side)
axhigh.set_title("Highest Day\n" + highest_day.strftime("%m/%d/%Y"))
ind = np.argmin(totals)
lowest_day = new_times[ind][0]
left_side = np.argmax(times == new_times[ind][0])#hackish, but oh well
right_side = np.argmax(times == new_times[ind][-1])#hackish, but oh well
make_interval_plot(d, axlow, left_side, right_side)
axlow.set_title("Lowest Day\n" + lowest_day.strftime("%m/%d/%Y"))
def gen_over_thresh(d, thresh):
"""A generatore that yields the times where the energy usage was above some threshold.
Parameters:
d -- The building record.
thresh -- The threshold.
Yields a pair of indices -- the times on either side of the point which crosses the threshold.
(12 hours on either side).
"""
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
temps, temps_oriflag = d["temps"]
#Right now, just assume you start and end below thresh
left_sides = [ind for ind in range(len(kwhs)-1) if kwhs[ind] <= thresh and kwhs[ind+1] > thresh]
right_sides = [ind for ind in range(len(kwhs)-1) if kwhs[ind] > thresh and kwhs[ind+1] <= thresh]
#Now account for the fact that you may start/end above thresh
if kwhs[0] >= thresh:
left_sides = [0] + left_sides
if kwhs[-1] >= thresh:
right_sides += [len(kwhs)-1]
periods = zip(left_sides, right_sides)
for ls, rs in periods:
new_left_side = max(0, ls - 12)
new_right_side = min(len(kwhs)-1, rs + 12)
kvals = kwhs[new_left_side:new_right_side]
ptimes = times[new_left_side:new_right_side]
tvals = temps[new_left_side:new_right_side]
yield new_left_side, new_right_side
def get_times_of_highest_change(d, num_times, direction = "increase"):
"""Returns the indices where the highest increase in electricity usage occured.
Note the indices returned are for just before the spikes occured (as opposed to just after).
Parameters:
d -- The building record.
num_times -- The number of times to be returned.
direction -- Either "increase" (default) or "decrease"
"""
kwhs, kwhs_oriflag = d["kwhs"]
times = d["times"]
first_deriv = kwhs[1:] - kwhs[:-1]
inds = np.argsort(first_deriv)
if direction == "increase":
inds = inds[-num_times:][::-1]
else:
inds = inds[:num_times]
return inds
def make_interval_plot(d, ax, start, end, show_temps = True, show_sun = True, c = 'blue'):
"""
Given a building record, an axis, and a start/end, plot the interval between the start and end.
Parameters:
d -- The building record.
ax -- The axis.
start -- Index of start time.
end -- Index of end time.
show_temps [True] -- If True, temperatures are shown in red.
show_sun [True] -- If True, sunlight is shown in dashed grey.
c ['blue'] -- The color of the kwhs line.
"""
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
temps, temps_oriflag = d["temps"]
kwhs = kwhs[start:end]
kwhs_oriflag = kwhs_oriflag[start:end]
temps = temps[start:end]
temps_oriflag = temps_oriflag[start:end]
times = times[start:end]
lns1 = ax.plot(times, kwhs, label = "kwhs", c = c)
ax.set_ylabel("kwh")
suns = np.array([max(0, getSun("IL", x.replace(tzinfo = tz_used))) for x in times])
sun_ax = ax.twinx()
lns2 = sun_ax.plot(times, suns, label = "Sunlight", c = "purple", alpha = 0.3, ls = "dashed")
sun_ax.set_ylim((-.5, 1))
sun_ax.set_yticks([])
weather_ax = sun_ax.twinx()
weather_ax.set_ylabel("Temperature")
lns3 = weather_ax.plot(times, temps, label = "Temperature", c = "red", alpha = 0.4)
lns = lns1+lns2+lns3
labs = [l.get_label() for l in lns]
weather_ax.legend(lns, labs, loc=0)
ax.grid(True)
labels = ax.get_xticklabels()
for label in labels:
label.set_rotation(30)
def make_cami_fig(d, ax):
"""
Given a building record and an axis, add Cami's favorite plot to the axis.
Cami's favorite plot provides a detailed illustration of how much money can be saved.
This figure shows a variety of retrofitting options, and the predicted savings induced by each (as well as confidence intervals).
"""
num_hours = 24
kwhs, kwhs_oriflag = d["kwhs"]
times = d["times"]
is_sunday_start = (lambda x: x.weekday() == 6 and x.hour == 0)
weeks, new_times = get_periods(d, num_hours, is_sunday_start)
thresh = np.percentile(weeks, 95, axis = 0)
avg_week = np.average(weeks, axis = 0)
high_avgs = []
for col in range(num_hours):
inds = weeks[:, col] >thresh[col]
high_avgs.append(np.average(weeks[:, col][inds]))
high_avgs = np.array(high_avgs)
sum_avg = np.sum(avg_week)
sum_high = np.sum(high_avgs)
ax.set_title("Cami could have saved you:\n%.2f dollars" % ((sum_high - sum_avg) *.05))
ax.stackplot(range(num_hours), avg_week, high_avgs-avg_week, colors = ["white", "red"])
ax.plot(range(num_hours), thresh, ls = "dotted", color = "black")
def make_cluster_fig(d, types_ax, times_ax):
"""
Given a building record and two axes, this function populates the axes.
The first axis (types_ax) is pouplated with the means of the clusters found via clustering the individual days based on behavior.
The second axis (times_ax) is populated with the original signal over the full year, where each day is colored based on its found cluster.
"""
kwhs, kwhs_oriflag = d["kwhs"]
times = d["times"]
is_midnight = (lambda x: x.hour == 0)
days, new_times = get_periods(d, 24, is_midnight)
oridays = copy.copy(days)
times_ax.plot(times, kwhs, lw=0.5, c="black")
for day in days:
day -= np.average(day)#center each day
num_clusters = 3
clusterer = KMeans(init='k-means++', n_clusters=num_clusters, n_init=10)
#clusterer = mixture.GMM(n_components=3, covariance_type='full')
#clusterer = mixture.DPGMM(n_components=3, covariance_type='full')
clusterer.fit(days)
preds = clusterer.predict(days)
#centers = clusterer.means_
centers = clusterer.cluster_centers_
num_in_each = []
for c in range(len(centers)):
num_in_each.append(len(preds[preds == c]))
sorted_inds = [x[1] for x in sorted(zip(num_in_each, range(len(num_in_each))), reverse = True)]
colors = ["green", "blue", "red", "purple", "pink", "black", "grey"]
colors = colors[:len(centers)]
cmap = dict(zip(sorted_inds, colors))
for c, center in enumerate(centers):
types_ax.plot(center, label = "# days: " + str(len(preds[preds == c])), c = cmap[c])
types_ax.legend()
types_ax.set_yticks([])
types_ax.set_ylabel("Relative energy usage")
for i, d in enumerate(oridays):
times_ax.plot(new_times[i], d, c = cmap[preds[i]])
times_ax.set_ylabel("kwhs")
def make_deriv_day_fig(d, ax):
is_midnight = (lambda x: x.hour == 0)
times = d["times"]
kwhs, kwhs_oriflag = d["kwhs"]
deriv = kwhs[1:] - kwhs[:-1]
hods = np.array([t.hour for t in times[:-1]])
trim = True
if trim:
#remove 1% of data (extreme values)
upper = np.percentile(deriv, 99.5)
lower = np.percentile(deriv, 0.5)
flag = np.array([lower < x and x < upper for x in deriv])
deriv = deriv[flag]
hods = hods[flag]
ax.hist2d(hods, deriv, bins = (23*3, 100), norm = LogNorm())
ax.set_xticks(range(23))
ax.axhline(0, 0, 1, ls = "dotted", c = "black")
ax.set_xlabel("Hours after midnight")
ax.set_ylabel("Change in kwh")
def multi_plot(d, foutn = None):
fontsize = 24
if foutn == None:
foutn = fig_loc + 'fin_' + str(d["bid"]) + '_' + str(the_year) + '.pdf'
pdf = PdfPages(foutn)
size = (8.5, 11)
#size = (13.6, 7.7)
all_figs = ["general",
"avg behavior",
"clustering",
"box plots",
"behavior",
"spikes",
"deriv day",
"outliers2",
"holidays",
"overthresh",
"extreme days",
"raw"]
bmax = len(all_figs)
for i, f in enumerate(all_figs):
add_fig(pdf, d, f, size = size, fontsize = fontsize)
progress_bar(i+1, bmax)
print "\n",
pdf.close()
def _add_fig_box_plots(pdf, d, size, fontsize):
#Box-plot figures
b_fig = plt.figure(figsize = size)
ax_alldays = b_fig.add_subplot(211)
ax_split = b_fig.add_subplot(212)
make_boxplot_all_days_fig(d, ax_alldays)
make_boxplot_weekday_vs_end_fig(d, ax_split)
b_fig.suptitle("Effect of Schedule", fontsize = fontsize)
plt.savefig(pdf, format = 'pdf')
def _add_fig_general(pdf, d, size, fontsize):
#General/global figure
g_fig = plt.figure(figsize = size)
g_text = g_fig.add_subplot(1, 2, 1)
g_hist = g_fig.add_subplot(2, 2, 2)
g_totals = g_fig.add_subplot(2, 2, 4)
# make_text_fig(d, g_text)
make_hist_fig(d, g_hist)
make_monthly_usage_fig(d, g_totals)
bid = d["bid"]
try:
feature_map, desc = qload("prison_features_map.pkl")
if isinstance(bid, str) and "q" in bid:
features = feature_map[int(bid[:-2])]
else:
features = feature_map[bid]
g_fig.suptitle(features["name"], fontsize = fontsize)
make_prison_text_fig(d, g_text)
except:
g_fig.suptitle(str(bid), fontsize = fontsize)
make_text_fig(d, g_text)
plt.savefig(pdf, format = 'pdf')
def _add_fig_avg_behavior(pdf, d, size, fontsize):
#Avg behavior figure
n_fig = plt.figure(figsize = size)
n_avgday = n_fig.add_subplot(2, 1, 1)
n_avgweek = n_fig.add_subplot(2, 1, 2)
make_avg_day_fig(d, n_avgday)
make_avg_week_fig(d, n_avgweek)
n_fig.suptitle("Average Behavior", fontsize = fontsize)
extract_legend(n_fig)
plt.savefig(pdf, format = 'pdf')
def _add_fig_behavior(pdf, d, size, fontsize):
#Behavior fig
b_fig = plt.figure(figsize = size)
b_vstemp = b_fig.add_subplot(2, 2, 1)
b_vssun = b_fig.add_subplot(2, 2, 3)
b_vstempad = b_fig.add_subplot(2, 2, 2)
b_vssunad = b_fig.add_subplot(2, 2, 4)
make_temp_vs_kwh_fig(d, b_vstemp, False)
make_kwh_vs_sun_fig(d, b_vssun)
make_temp_vs_kwh_fig(d, b_vstempad, True)
make_kwh_vs_sun_fig(d, b_vssunad, True)
b_fig.suptitle("Average Behavior", fontsize = fontsize)
extract_legend(b_fig)
plt.savefig(pdf, format = 'pdf')
def _add_fig_raw(pdf, d, size, fontsize):
#Raw data figure
a_fig = plt.figure(figsize = size)
a_temps = a_fig.add_subplot(3, 1, 1)
a_kwhs = a_fig.add_subplot(3, 1, 2)
a_freqs = a_fig.add_subplot(3, 1, 3)
make_temp_vs_time_fig(d, a_temps)
make_kwhs_vs_time_fig(d, a_kwhs)
make_freqs_fig(d, a_freqs)
a_fig.suptitle("Raw Data", fontsize = fontsize)
extract_legend(a_fig)
plt.savefig(pdf, format = 'pdf')
def _add_fig_outliers(pdf, d, size, fontsize):
#outliers
o_fig = plt.figure(figsize = size)
avg_day = o_fig.add_subplot(5, 2, 1)
make_avg_day_fig(d, avg_day)
times = d["times"]
days = gen_strange_pers(d, 3, period = "day")
for i, p in enumerate(days):
day_fig = o_fig.add_subplot(4, 2, 2*(i + 2)-1)
make_strange_per_fig(d, day_fig, p)
day_fig.set_title(times[p[0]].strftime("%m/%d/%y"))
avg_week = o_fig.add_subplot(5, 2, 2)
make_avg_week_fig(d, avg_week)
weeks = gen_strange_pers(d, 3, period = "week")
for i, p in enumerate(weeks):
week_fig = o_fig.add_subplot(4, 2, 2*(i + 2))
make_strange_per_fig(d, week_fig, p)
week_fig.set_title(times[p[0]].strftime("%m/%d/%y") + " - " +
times[p[1]].strftime("%m/%d/%y"))
o_fig.suptitle("Outliers", fontsize = fontsize)
extract_legend(o_fig)
plt.subplots_adjust(hspace = .55)
plt.savefig(pdf, format = 'pdf')
def _add_fig_outliers2(pdf, d, size, fontsize):
#outliers, 2 page
times = d["times"]