forked from h2oai/h2o-2
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathClaimsTweedieAnalysis_8.2.twb
1699 lines (1698 loc) · 130 KB
/
ClaimsTweedieAnalysis_8.2.twb
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
<?xml version='1.0' encoding='utf-8' ?>
<workbook source-platform='win' version='8.3' xmlns:user='http://www.tableausoftware.com/xml/user'>
<!-- build 8200.14.0819.2015 -->
<preferences>
</preferences>
<datasources>
<datasource hasconnection='false' inline='true' name='Parameters' version='8.3'>
<aliases enabled='yes' />
<column caption='H2O Port' datatype='integer' name='[Parameter 1]' param-domain-type='any' role='measure' type='quantitative' value='54321'>
<calculation class='tableau' formula='54321' />
</column>
<column caption='H2O IP' datatype='string' name='[Parameter 2]' param-domain-type='any' role='measure' type='nominal' value='"localhost"'>
<calculation class='tableau' formula='"localhost"' />
</column>
<column caption='H2O Max Mem Size' datatype='string' name='[Parameter 3]' param-domain-type='any' role='measure' type='nominal' value='"4g"'>
<calculation class='tableau' formula='"4g"' />
</column>
<column caption='Claims Data File' datatype='string' name='[Parameter 4]' param-domain-type='any' role='measure' type='nominal' value='"d:\\claimsdata.csv"'>
<calculation class='tableau' formula='"d:\claimsdata.csv"' />
</column>
</datasource>
<datasource caption='claims_coefficients' inline='true' name='textscan.41886.636795671293' version='8.3'>
<connection class='textscan' directory='D:/' filename='claims_coefficients.csv'>
<relation name='claims_coefficients#csv' table='[claims_coefficients#csv]' type='table'>
<columns character-set='UTF-8' header='yes' locale='en_US'>
<column datatype='string' name='Coefficients' ordinal='0' />
</columns>
</relation>
<metadata-records>
<metadata-record class='column'>
<remote-name>Coefficients</remote-name>
<remote-type>129</remote-type>
<local-name>[Coefficients]</local-name>
<parent-name>[claims_coefficients#csv]</parent-name>
<remote-alias>Coefficients</remote-alias>
<ordinal>0</ordinal>
<local-type>string</local-type>
<aggregation>Count</aggregation>
<scale>1</scale>
<width>1073741823</width>
<contains-null>true</contains-null>
<collation flag='0' name='LEN_RUS' />
<attributes>
<attribute datatype='string' name='DebugRemoteCollation'>"en_US"</attribute>
<attribute datatype='string' name='DebugRemoteMetadata (compression)'>"heap"</attribute>
<attribute datatype='integer' name='DebugRemoteMetadata (size)'>4294967292</attribute>
<attribute datatype='integer' name='DebugRemoteMetadata (storagewidth)'>8</attribute>
<attribute datatype='string' name='DebugRemoteType'>"str"</attribute>
</attributes>
</metadata-record>
<metadata-record class='capability'>
<remote-name></remote-name>
<remote-type>0</remote-type>
<parent-name>[claims_coefficients#csv]</parent-name>
<remote-alias></remote-alias>
<aggregation>Count</aggregation>
<contains-null>true</contains-null>
<attributes>
<attribute datatype='string' name='character-set'>"UTF-8"</attribute>
<attribute datatype='string' name='header-row'>"true"</attribute>
<attribute datatype='string' name='locale'>"en_US"</attribute>
<attribute datatype='string' name='single-char'>""</attribute>
</attributes>
</metadata-record>
</metadata-records>
</connection>
<aliases enabled='yes' />
<column caption='Coefficient Name' datatype='string' name='[Coefficients]' role='dimension' type='nominal'>
</column>
<column caption='Normalized Value' datatype='real' name='[Number of Claims Filed (copy) 1]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' tweedie.vi[.arg1] ', MAX([Coefficients]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column datatype='integer' name='[Number of Records]' role='measure' type='quantitative' user:auto-column='numrec'>
<calculation class='tableau' formula='1' />
</column>
<layout dim-ordering='alphabetic' dim-percentage='0.472789' measure-ordering='alphabetic' measure-percentage='0.309524' show-structure='true' />
<semantic-values>
<semantic-value key='[Country].[Name]' value='"United States"' />
</semantic-values>
</datasource>
<datasource caption='claims_metadata' inline='true' name='textscan.41885.651995601853' version='8.3'>
<connection class='textscan' directory='D:/' filename='claims_losscodes.csv'>
<relation name='claims_metadata#csv' table='[claims_metadata#csv]' type='table'>
<columns character-set='UTF-8' header='yes' locale='en_US' separator=','>
<column datatype='date' name='MonthlyDate' ordinal='0' />
<column datatype='string' name='StateName' ordinal='1' />
<column datatype='string' name='LossCode' ordinal='2' />
</columns>
</relation>
<metadata-records>
<metadata-record class='column'>
<remote-name>MonthlyDate</remote-name>
<remote-type>133</remote-type>
<local-name>[MonthlyDate]</local-name>
<parent-name>[claims_metadata#csv]</parent-name>
<remote-alias>MonthlyDate</remote-alias>
<ordinal>0</ordinal>
<local-type>date</local-type>
<aggregation>Year</aggregation>
<contains-null>true</contains-null>
<attributes>
<attribute datatype='integer' name='DebugRemoteMetadata (size)'>4</attribute>
<attribute datatype='string' name='DebugRemoteType'>"date"</attribute>
</attributes>
</metadata-record>
<metadata-record class='column'>
<remote-name>StateName</remote-name>
<remote-type>129</remote-type>
<local-name>[StateName]</local-name>
<parent-name>[claims_metadata#csv]</parent-name>
<remote-alias>StateName</remote-alias>
<ordinal>1</ordinal>
<local-type>string</local-type>
<aggregation>Count</aggregation>
<scale>1</scale>
<width>1073741823</width>
<contains-null>true</contains-null>
<collation flag='0' name='LEN_RUS' />
<attributes>
<attribute datatype='string' name='DebugRemoteCollation'>"en_US"</attribute>
<attribute datatype='string' name='DebugRemoteMetadata (compression)'>"heap"</attribute>
<attribute datatype='integer' name='DebugRemoteMetadata (size)'>4294967292</attribute>
<attribute datatype='integer' name='DebugRemoteMetadata (storagewidth)'>8</attribute>
<attribute datatype='string' name='DebugRemoteType'>"str"</attribute>
</attributes>
</metadata-record>
<metadata-record class='column'>
<remote-name>LossCode</remote-name>
<remote-type>129</remote-type>
<local-name>[LossCode]</local-name>
<parent-name>[claims_metadata#csv]</parent-name>
<remote-alias>LossCode</remote-alias>
<ordinal>2</ordinal>
<local-type>string</local-type>
<aggregation>Count</aggregation>
<scale>1</scale>
<width>1073741823</width>
<contains-null>true</contains-null>
<collation flag='0' name='LEN_RUS' />
<attributes>
<attribute datatype='string' name='DebugRemoteCollation'>"en_US"</attribute>
<attribute datatype='string' name='DebugRemoteMetadata (compression)'>"heap"</attribute>
<attribute datatype='integer' name='DebugRemoteMetadata (size)'>4294967292</attribute>
<attribute datatype='integer' name='DebugRemoteMetadata (storagewidth)'>8</attribute>
<attribute datatype='string' name='DebugRemoteType'>"str"</attribute>
</attributes>
</metadata-record>
<metadata-record class='capability'>
<remote-name></remote-name>
<remote-type>0</remote-type>
<parent-name>[claims_metadata#csv]</parent-name>
<remote-alias></remote-alias>
<aggregation>Count</aggregation>
<contains-null>true</contains-null>
<attributes>
<attribute datatype='string' name='character-set'>"UTF-8"</attribute>
<attribute datatype='string' name='field-delimiter'>","</attribute>
<attribute datatype='string' name='header-row'>"true"</attribute>
<attribute datatype='string' name='locale'>"en_US"</attribute>
<attribute datatype='string' name='single-char'>""</attribute>
</attributes>
</metadata-record>
</metadata-records>
</connection>
<aliases enabled='yes' />
<column caption='H2O Initialization' datatype='boolean' name='[Calculation_1010902082708389]' role='measure' type='nominal'>
<calculation class='tableau' formula='SCRIPT_BOOL( ' myh2oCleaner <- function(pattern = "Last.value.") { keys <- h2o.ls(h2oHandle, pattern = pattern)$Key if (!is.null(keys)) h2o.rm(h2oHandle, keys) invisible(keys) } if (!exists("h2oHandle", globalenv())) { library(h2o) h2oHandle <- h2o.init(ip = .arg1, port = .arg2, max_mem_size = .arg3, nthreads = -1) } if (length(h2o.ls(h2oHandle, pattern = "claims.dat")) == 0L) { claims.dat <- h2o.importFile(h2oHandle, path = .arg4, key = "claims.dat") # Create indicator variables for claim payout percentages invisible( h2o.exec(claims.dat$ClaimPaidRatio <- claims.dat$TotalPaid / claims.dat$TotalClaim) ) payoutThreshold <- 1 invisible( h2o.exec(claims.dat$HighClaimPaid <- claims.dat$ClaimPaidRatio > payoutThreshold) ) # Create log transformed variables invisible( h2o.exec(claims.dat$logDeduct <- log(claims.dat$Deduct + 1)) ) invisible( h2o.exec(claims.dat$logFees <- log(claims.dat$Fees + 1)) ) invisible( h2o.exec(claims.dat$logTotalClaim <- log(claims.dat$TotalClaim + 1)) ) invisible( h2o.exec(claims.dat$NonNegPaid <- ifelse(claims.dat$TotalPaid < 0, 0, claims.dat$TotalPaid)) ) invisible( h2o.exec(claims.dat$logTotalPaid <- log(claims.dat$NonNegPaid + 1)) ) # Convert some numeric columns to factor for (j in c("CallCenterPostalCode")) claims.dat[[j]] <- as.factor(claims.dat[[j]]) myh2oCleaner() # Limit analysis to "Claim Complete" completed.claims.dat <- claims.dat[claims.dat$ResponseStatus == "Claim Complete", ] completed.claims.dat <- h2o.assign(completed.claims.dat, "completed.claims.dat") myh2oCleaner() xvars <- c("CallCenterPostalCode", "Demographics_AgeGroup", "Demographics_Gender", "FirstClaim", "INStype", "LossCode", "MonthlyDate", "SourceCode", "StateName", "logDeduct", "logFees", "logTotalClaim") tweedie.final <- h2o.glm(xvars, "NonNegPaid", completed.claims.dat, key = "tweedie.final", family = "tweedie", tweedie.p = 1.75, alpha = 0, lambda = 12, variable_importances = TRUE) tweedie.vi <- sort(abs(tweedie.final@model$normalized_coefficients), decreasing = TRUE) tweedie.vi <- tweedie.vi[names(tweedie.vi) != "Intercept"] myh2oCleaner(pattern = "GLM2Job__") completed.claims.dat$predicted <- h2o.predict(tweedie.final, completed.claims.dat) myh2oCleaner(pattern = "GLM2Predict_") library(tweedie) completed.claims.dat$deviance <- tweedie.dev(y = completed.claims.dat$NonNegPaid, mu = completed.claims.dat$predicted, power = 1.75) claims.tweediedev.dat <- completed.claims.dat[!is.na(completed.claims.dat$deviance), ] claims.tweediedev.dat <- h2o.assign(claims.tweediedev.dat, "claims.tweediedev.dat") myh2oCleaner() devAggrFun <- function(x) { cbind(N = nrow(x), SumTotalPaid = sum(x$TotalPaid), SumPredicted = sum(x$predicted), MinDeviance = min(x$deviance), MeanDeviance = mean(x$deviance), MaxDeviance = max(x$deviance)) } h2o.addFunction(h2oHandle, devAggrFun) devAggr <- as.data.frame(h2o.ddply(claims.tweediedev.dat, c("MonthlyDate", "StateName", "LossCode"), devAggrFun)) devAggr[[1L]] <- as.Date(as.character(devAggr[[1L]]), "%m/%d/%Y") devAggr[[4L]] <- as.integer(devAggr[[4L]]) colnames(devAggr)[-(1:3)] <- c("N", "SumTotalPaid", "SumPredicted", "MinDeviance", "MeanDeviance", "MaxDeviance") print(head(devAggr, 10)) myh2oCleaner() } TRUE ', [Parameters].[Parameter 2], [Parameters].[Parameter 1], [Parameters].[Parameter 3], [Parameters].[Parameter 4])' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Loss Code' datatype='string' name='[LossCode]' role='dimension' type='nominal'>
</column>
<column caption='Min Deviance' datatype='real' name='[Max Deviance (copy 2)]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "MinDeviance", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Number of Completed Claims' datatype='integer' name='[Max Deviance (copy)]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_INT( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_integer_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "N", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Actual Payout' datatype='real' name='[Mean Deviance (copy)]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "SumTotalPaid", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Claim Date' datatype='date' name='[MonthlyDate]' role='dimension' type='ordinal'>
</column>
<column caption='Max Deviance' datatype='real' name='[Number of Claims Filed (copy) 1 1]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "MaxDeviance", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Mean Deviance' datatype='real' name='[Number of Claims Filed (copy) 1]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "MeanDeviance", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column datatype='integer' name='[Number of Records]' role='measure' type='quantitative' user:auto-column='numrec'>
<calculation class='tableau' formula='1' />
</column>
<column caption='State Name' datatype='string' name='[StateName]' role='dimension' semantic-role='[State].[Name]' type='nominal'>
</column>
<column caption='Predicted Payout' datatype='real' name='[Sum Total Paid (copy)]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "SumPredicted", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column-instance column='[LossCode]' derivation='Attribute' name='[attr:LossCode:nk]' pivot='key' type='nominal' />
<column-instance column='[LossCode]' derivation='None' name='[none:LossCode:nk]' pivot='key' type='nominal' />
<column-instance column='[Mean Deviance (copy)]' derivation='User' name='[usr:Mean Deviance (copy):qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
<column-instance column='[Sum Total Paid (copy)]' derivation='User' name='[usr:Sum Total Paid (copy):qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
<group caption='Action (Loss Code)' hidden='true' name='[Action (Loss Code)]' name-style='unqualified' user:auto-column='sheet_link'>
<groupfilter function='crossjoin'>
<groupfilter function='level-members' level='[LossCode]' />
</groupfilter>
</group>
<group caption='Action (State Name)' hidden='true' name='[Action (State Name)]' name-style='unqualified' user:auto-column='sheet_link'>
<groupfilter function='crossjoin'>
<groupfilter function='level-members' level='[StateName]' />
</groupfilter>
</group>
<layout dim-ordering='alphabetic' dim-percentage='0.472789' measure-ordering='alphabetic' measure-percentage='0.282313' show-structure='true' />
<style>
<style-rule element='mark'>
<encoding attr='color' field='[attr:LossCode:nk]' type='palette'>
<map to='#17becf'>
<bucket>"medlr"</bucket>
</map>
<map to='#1f77b4'>
<bucket>"accer"</bucket>
</map>
<map to='#1f77b4'>
<bucket>"nglhr"</bucket>
</map>
<map to='#2ca02c'>
<bucket>"dep"</bucket>
</map>
<map to='#2ca02c'>
<bucket>"pulb"</bucket>
</map>
<map to='#7f7f7f'>
<bucket>"hemn"</bucket>
</map>
<map to='#8c564b'>
<bucket>"flr"</bucket>
</map>
<map to='#8c564b'>
<bucket>"tort"</bucket>
</map>
<map to='#9467bd'>
<bucket>"er-mt"</bucket>
</map>
<map to='#9467bd'>
<bucket>"suit"</bucket>
</map>
<map to='#98df8a'>
<bucket>"emt"</bucket>
</map>
<map to='#98df8a'>
<bucket>"pulhr"</bucket>
</map>
<map to='#9edae5'>
<bucket>"medpa"</bucket>
</map>
<map to='#aec7e8'>
<bucket>"carr"</bucket>
</map>
<map to='#aec7e8'>
<bucket>"nurf"</bucket>
</map>
<map to='#bcbd22'>
<bucket>"inf"</bucket>
</map>
<map to='#c49c94'>
<bucket>"flrf"</bucket>
</map>
<map to='#c5b0d5'>
<bucket>"er-x"</bucket>
</map>
<map to='#c5b0d5'>
<bucket>"surmp"</bucket>
</map>
<map to='#c7c7c7'>
<bucket>"impay"</bucket>
</map>
<map to='#d62728'>
<bucket>"er-"</bucket>
</map>
<map to='#d62728'>
<bucket>"revm"</bucket>
</map>
<map to='#dbdb8d'>
<bucket>"medh"</bucket>
</map>
<map to='#e377c2'>
<bucket>"flron"</bucket>
</map>
<map to='#f7b6d2'>
<bucket>"flrst"</bucket>
</map>
<map to='#ff7f0e'>
<bucket>"cheh"</bucket>
</map>
<map to='#ff7f0e'>
<bucket>"obgor"</bucket>
</map>
<map to='#ff9896'>
<bucket>"er-ag"</bucket>
</map>
<map to='#ff9896'>
<bucket>"sui"</bucket>
</map>
<map to='#ffbb78'>
<bucket>"dc-"</bucket>
</map>
<map to='#ffbb78'>
<bucket>"oth"</bucket>
</map>
</encoding>
<encoding attr='color' field='[none:LossCode:nk]' type='palette'>
<map to='#17becf'>
<bucket>"medlr"</bucket>
</map>
<map to='#1f77b4'>
<bucket>"accer"</bucket>
</map>
<map to='#1f77b4'>
<bucket>"nglhr"</bucket>
</map>
<map to='#2ca02c'>
<bucket>"dep"</bucket>
</map>
<map to='#2ca02c'>
<bucket>"pulb"</bucket>
</map>
<map to='#7f7f7f'>
<bucket>"hemn"</bucket>
</map>
<map to='#8c564b'>
<bucket>"flr"</bucket>
</map>
<map to='#8c564b'>
<bucket>"tort"</bucket>
</map>
<map to='#9467bd'>
<bucket>"er-mt"</bucket>
</map>
<map to='#9467bd'>
<bucket>"suit"</bucket>
</map>
<map to='#98df8a'>
<bucket>"emt"</bucket>
</map>
<map to='#98df8a'>
<bucket>"pulhr"</bucket>
</map>
<map to='#9edae5'>
<bucket>"medpa"</bucket>
</map>
<map to='#aec7e8'>
<bucket>"carr"</bucket>
</map>
<map to='#aec7e8'>
<bucket>"nurf"</bucket>
</map>
<map to='#bcbd22'>
<bucket>"inf"</bucket>
</map>
<map to='#c49c94'>
<bucket>"flrf"</bucket>
</map>
<map to='#c5b0d5'>
<bucket>"er-x"</bucket>
</map>
<map to='#c5b0d5'>
<bucket>"surmp"</bucket>
</map>
<map to='#c7c7c7'>
<bucket>"impay"</bucket>
</map>
<map to='#d62728'>
<bucket>"er-"</bucket>
</map>
<map to='#d62728'>
<bucket>"revm"</bucket>
</map>
<map to='#dbdb8d'>
<bucket>"medh"</bucket>
</map>
<map to='#e377c2'>
<bucket>"flron"</bucket>
</map>
<map to='#f7b6d2'>
<bucket>"flrst"</bucket>
</map>
<map to='#ff7f0e'>
<bucket>"cheh"</bucket>
</map>
<map to='#ff7f0e'>
<bucket>"obgor"</bucket>
</map>
<map to='#ff9896'>
<bucket>"er-ag"</bucket>
</map>
<map to='#ff9896'>
<bucket>"sui"</bucket>
</map>
<map to='#ffbb78'>
<bucket>"dc-"</bucket>
</map>
<map to='#ffbb78'>
<bucket>"oth"</bucket>
</map>
</encoding>
<encoding attr='color' field='[:Measure Names]' type='palette'>
<map to='#1f77b4'>
<bucket>"[textscan.41885.651995601853].[usr:Sum Total Paid (copy):qk]"</bucket>
</map>
<map to='#d62728'>
<bucket>"[textscan.41885.651995601853].[usr:Mean Deviance (copy):qk]"</bucket>
</map>
</encoding>
</style-rule>
</style>
<semantic-values>
<semantic-value key='[Country].[Name]' value='"United States"' />
</semantic-values>
<datasource-dependencies datasource='Parameters'>
<column caption='H2O Port' datatype='integer' name='[Parameter 1]' param-domain-type='any' role='measure' type='quantitative' value='54321'>
<calculation class='tableau' formula='54321' />
</column>
<column caption='H2O IP' datatype='string' name='[Parameter 2]' param-domain-type='any' role='measure' type='nominal' value='"localhost"'>
<calculation class='tableau' formula='"localhost"' />
</column>
<column caption='H2O Max Mem Size' datatype='string' name='[Parameter 3]' param-domain-type='any' role='measure' type='nominal' value='"4g"'>
<calculation class='tableau' formula='"4g"' />
</column>
<column caption='Claims Data File' datatype='string' name='[Parameter 4]' param-domain-type='any' role='measure' type='nominal' value='"d:\\claimsdata.csv"'>
<calculation class='tableau' formula='"d:\claimsdata.csv"' />
</column>
</datasource-dependencies>
</datasource>
</datasources>
<mapsources>
<mapsource name='Tableau' />
</mapsources>
<actions>
<action caption='State Name' name='[Action1]'>
<activation type='on-select' />
<source dashboard='Predicted vs Actual' type='sheet' worksheet='Maximum Deviance' />
<link caption='State Name' delimiter=',' escape='\' expression='tsl:Predicted%20vs%20Actual?%5Btextscan.41885.651995601853%5D.%5BStateName%5D~s0=<[StateName]~na>' include-null='true' multi-select='true' url-escape='true' />
<command command='tsc:tsl-filter'>
<param name='exclude' value='Maximum Deviance' />
<param name='target' value='Predicted vs Actual' />
</command>
</action>
<action caption='Filter by Loss Code' name='[Action2]'>
<activation type='on-select' />
<source dashboard='Predicted vs Actual' type='sheet' worksheet='Payments by Loss Code' />
<link caption='Filter by Loss Code' delimiter=',' escape='\' expression='tsl:Predicted%20vs%20Actual?%5Btextscan.41885.651995601853%5D.%5BLossCode%5D~s0=<[LossCode]~na>' include-null='true' multi-select='true' url-escape='true' />
<command command='tsc:tsl-filter'>
<param name='exclude' value='Payments by Loss Code' />
<param name='single-select' value='' />
<param name='special-fields' value='all' />
<param name='target' value='Predicted vs Actual' />
</command>
</action>
</actions>
<worksheets>
<worksheet name='Maximum Deviance'>
<layout-options>
<title>
<formatted-text>
<run fontsize='13'><</run>
<run fontsize='13'>Sheet Name</run>
<run fontsize='13'>></run>
<run fontsize='13'> for Loss Code </run>
<run fontsize='13'><</run>
<run fontsize='13'>[textscan.41885.651995601853].[attr:LossCode:nk]</run>
<run fontsize='13'>></run>
</formatted-text>
</title>
</layout-options>
<table>
<view>
<datasources>
<datasource caption='claims_metadata' name='textscan.41885.651995601853' />
<datasource name='Parameters' />
</datasources>
<mapsources>
<mapsource name='Tableau' />
</mapsources>
<datasource-dependencies datasource='textscan.41885.651995601853'>
<column caption='H2O Initialization' datatype='boolean' name='[Calculation_1010902082708389]' role='measure' type='nominal'>
<calculation class='tableau' formula='SCRIPT_BOOL( ' myh2oCleaner <- function(pattern = "Last.value.") { keys <- h2o.ls(h2oHandle, pattern = pattern)$Key if (!is.null(keys)) h2o.rm(h2oHandle, keys) invisible(keys) } if (!exists("h2oHandle", globalenv())) { library(h2o) h2oHandle <- h2o.init(ip = .arg1, port = .arg2, max_mem_size = .arg3, nthreads = -1) } if (length(h2o.ls(h2oHandle, pattern = "claims.dat")) == 0L) { claims.dat <- h2o.importFile(h2oHandle, path = .arg4, key = "claims.dat") # Create indicator variables for claim payout percentages invisible( h2o.exec(claims.dat$ClaimPaidRatio <- claims.dat$TotalPaid / claims.dat$TotalClaim) ) payoutThreshold <- 1 invisible( h2o.exec(claims.dat$HighClaimPaid <- claims.dat$ClaimPaidRatio > payoutThreshold) ) # Create log transformed variables invisible( h2o.exec(claims.dat$logDeduct <- log(claims.dat$Deduct + 1)) ) invisible( h2o.exec(claims.dat$logFees <- log(claims.dat$Fees + 1)) ) invisible( h2o.exec(claims.dat$logTotalClaim <- log(claims.dat$TotalClaim + 1)) ) invisible( h2o.exec(claims.dat$NonNegPaid <- ifelse(claims.dat$TotalPaid < 0, 0, claims.dat$TotalPaid)) ) invisible( h2o.exec(claims.dat$logTotalPaid <- log(claims.dat$NonNegPaid + 1)) ) # Convert some numeric columns to factor for (j in c("CallCenterPostalCode")) claims.dat[[j]] <- as.factor(claims.dat[[j]]) myh2oCleaner() # Limit analysis to "Claim Complete" completed.claims.dat <- claims.dat[claims.dat$ResponseStatus == "Claim Complete", ] completed.claims.dat <- h2o.assign(completed.claims.dat, "completed.claims.dat") myh2oCleaner() xvars <- c("CallCenterPostalCode", "Demographics_AgeGroup", "Demographics_Gender", "FirstClaim", "INStype", "LossCode", "MonthlyDate", "SourceCode", "StateName", "logDeduct", "logFees", "logTotalClaim") tweedie.final <- h2o.glm(xvars, "NonNegPaid", completed.claims.dat, key = "tweedie.final", family = "tweedie", tweedie.p = 1.75, alpha = 0, lambda = 12, variable_importances = TRUE) tweedie.vi <- sort(abs(tweedie.final@model$normalized_coefficients), decreasing = TRUE) tweedie.vi <- tweedie.vi[names(tweedie.vi) != "Intercept"] myh2oCleaner(pattern = "GLM2Job__") completed.claims.dat$predicted <- h2o.predict(tweedie.final, completed.claims.dat) myh2oCleaner(pattern = "GLM2Predict_") library(tweedie) completed.claims.dat$deviance <- tweedie.dev(y = completed.claims.dat$NonNegPaid, mu = completed.claims.dat$predicted, power = 1.75) claims.tweediedev.dat <- completed.claims.dat[!is.na(completed.claims.dat$deviance), ] claims.tweediedev.dat <- h2o.assign(claims.tweediedev.dat, "claims.tweediedev.dat") myh2oCleaner() devAggrFun <- function(x) { cbind(N = nrow(x), SumTotalPaid = sum(x$TotalPaid), SumPredicted = sum(x$predicted), MinDeviance = min(x$deviance), MeanDeviance = mean(x$deviance), MaxDeviance = max(x$deviance)) } h2o.addFunction(h2oHandle, devAggrFun) devAggr <- as.data.frame(h2o.ddply(claims.tweediedev.dat, c("MonthlyDate", "StateName", "LossCode"), devAggrFun)) devAggr[[1L]] <- as.Date(as.character(devAggr[[1L]]), "%m/%d/%Y") devAggr[[4L]] <- as.integer(devAggr[[4L]]) colnames(devAggr)[-(1:3)] <- c("N", "SumTotalPaid", "SumPredicted", "MinDeviance", "MeanDeviance", "MaxDeviance") print(head(devAggr, 10)) myh2oCleaner() } TRUE ', [Parameters].[Parameter 2], [Parameters].[Parameter 1], [Parameters].[Parameter 3], [Parameters].[Parameter 4])' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Loss Code' datatype='string' name='[LossCode]' role='dimension' type='nominal'>
</column>
<column caption='Number of Completed Claims' datatype='integer' name='[Max Deviance (copy)]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_INT( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_integer_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "N", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Claim Date' datatype='date' name='[MonthlyDate]' role='dimension' type='ordinal'>
</column>
<column caption='Max Deviance' datatype='real' name='[Number of Claims Filed (copy) 1 1]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "MaxDeviance", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Mean Deviance' datatype='real' name='[Number of Claims Filed (copy) 1]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "MeanDeviance", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='State Name' datatype='string' name='[StateName]' role='dimension' semantic-role='[State].[Name]' type='nominal'>
</column>
<column-instance column='[LossCode]' derivation='Attribute' name='[attr:LossCode:nk]' pivot='key' type='nominal' />
<column-instance column='[MonthlyDate]' derivation='MY' name='[my:MonthlyDate:ok]' pivot='key' type='ordinal' />
<column-instance column='[LossCode]' derivation='None' name='[none:LossCode:nk]' pivot='key' type='nominal' />
<column-instance column='[StateName]' derivation='None' name='[none:StateName:nk]' pivot='key' type='nominal' />
<column-instance column='[Calculation_1010902082708389]' derivation='User' name='[usr:Calculation_1010902082708389:nk]' pivot='key' type='nominal'>
<table-calc ordering-type='Rows' />
</column-instance>
<column-instance column='[Max Deviance (copy)]' derivation='User' name='[usr:Max Deviance (copy):qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
<column-instance column='[Number of Claims Filed (copy) 1 1]' derivation='User' name='[usr:Number of Claims Filed (copy) 1 1:qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
<column-instance column='[Number of Claims Filed (copy) 1]' derivation='User' name='[usr:Number of Claims Filed (copy) 1:qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
</datasource-dependencies>
<filter class='categorical' column='[textscan.41885.651995601853].[Action (Loss Code)]'>
<groupfilter function='member' level='[LossCode]' member='"dep"' user:ui-action-filter='[Action2]' user:ui-domain='relevant' user:ui-enumeration='inclusive' user:ui-marker='enumerate' />
</filter>
<filter class='categorical' column='[textscan.41885.651995601853].[my:MonthlyDate:ok]'>
<groupfilter function='member' level='[my:MonthlyDate:ok]' member='201005' user:ui-domain='relevant' user:ui-enumeration='inclusive' user:ui-marker='enumerate' />
</filter>
<filter class='categorical' column='[textscan.41885.651995601853].[none:LossCode:nk]'>
<groupfilter function='level-members' level='[none:LossCode:nk]' user:ui-enumeration='all' user:ui-marker='enumerate' />
</filter>
<slices>
<column>[textscan.41885.651995601853].[my:MonthlyDate:ok]</column>
<column>[textscan.41885.651995601853].[none:LossCode:nk]</column>
<column>[textscan.41885.651995601853].[Action (Loss Code)]</column>
</slices>
<aggregation value='true' />
</view>
<style>
<style-rule element='mark'>
<encoding attr='color' field='[textscan.41885.651995601853].[usr:Number of Claims Filed (copy) 1 1:qk]' max='156.0' min='0.0' type='interpolated' />
</style-rule>
<style-rule element='worksheet'>
<format attr='in-tooltip' field='[textscan.41885.651995601853].[usr:Calculation_1010902082708389:nk]' value='false' />
</style-rule>
<style-rule element='map-layer'>
<format attr='enabled' id='tab_base' value='true' />
<format attr='enabled' id='um_lcover' value='true' />
<format attr='enabled' id='tab_coastline' value='false' />
<format attr='enabled' id='pp2_line' value='false' />
<format attr='enabled' id='light_adm0_bnd' value='false' />
<format attr='enabled' id='light_adm0_lbl' value='false' />
<format attr='enabled' id='um_adm0_bnd' value='true' />
<format attr='enabled' id='um_adm0_lbl' value='true' />
<format attr='enabled' id='light_pp2_statebounds' value='false' />
<format attr='enabled' id='light_pp2_statelabels' value='false' />
<format attr='enabled' id='pp2_adminlabels' value='false' />
<format attr='enabled' id='pp2_statebounds' value='true' />
<format attr='enabled' id='pp2_statelabels' value='true' />
<format attr='enabled' id='countybounds' value='false' />
<format attr='enabled' id='countylabels' value='false' />
<format attr='enabled' id='zipbounds' value='false' />
<format attr='enabled' id='ziplabels' value='false' />
<format attr='enabled' id='tab_areabounds' value='false' />
<format attr='enabled' id='tab_arealabels' value='false' />
<format attr='enabled' id='tab_msabounds' value='false' />
<format attr='enabled' id='tab_msalabels' value='false' />
</style-rule>
<style-rule element='map'>
<format attr='washout' value='0.0' />
</style-rule>
<style-rule element='quick-filter'>
<format attr='title' field='[textscan.41885.651995601853].[my:MonthlyDate:ok]' value='Claim Date'>
<formatted-text>
<run>Claim Date</run>
</formatted-text>
</format>
</style-rule>
</style>
<panes>
<pane>
<view>
<breakdown value='auto' />
</view>
<mark class='Automatic' />
<encodings>
<color column='[textscan.41885.651995601853].[usr:Number of Claims Filed (copy) 1 1:qk]' />
<lod column='[textscan.41885.651995601853].[none:StateName:nk]' />
<tooltip column='[textscan.41885.651995601853].[attr:LossCode:nk]' />
<lod column='[textscan.41885.651995601853].[usr:Calculation_1010902082708389:nk]' />
<tooltip column='[textscan.41885.651995601853].[usr:Max Deviance (copy):qk]' />
<tooltip column='[textscan.41885.651995601853].[usr:Number of Claims Filed (copy) 1:qk]' />
<geometry column='[textscan.41885.651995601853].[Geometry (generated)]' />
</encodings>
</pane>
</panes>
<rows>[textscan.41885.651995601853].[Latitude (generated)]</rows>
<cols>[textscan.41885.651995601853].[Longitude (generated)]</cols>
</table>
</worksheet>
<worksheet name='Payments by Loss Code'>
<layout-options>
<title>
<formatted-text>
<run fontsize='13'><</run>
<run fontsize='13'>[textscan.41885.651995601853].[my:MonthlyDate:ok]</run>
<run fontsize='13'>></run>
<run fontsize='13'>Æ </run>
<run fontsize='13'><</run>
<run fontsize='13'>Sheet Name</run>
<run fontsize='13'>></run>
<run fontsize='13'> for </run>
<run fontsize='13'><</run>
<run fontsize='13'>[textscan.41885.651995601853].[attr:StateName:nk]</run>
<run fontsize='13'>></run>
</formatted-text>
</title>
</layout-options>
<table>
<view>
<datasources>
<datasource caption='claims_metadata' name='textscan.41885.651995601853' />
</datasources>
<datasource-dependencies datasource='textscan.41885.651995601853'>
<column caption='Loss Code' datatype='string' name='[LossCode]' role='dimension' type='nominal'>
</column>
<column caption='Actual Payout' datatype='real' name='[Mean Deviance (copy)]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "SumTotalPaid", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column caption='Claim Date' datatype='date' name='[MonthlyDate]' role='dimension' type='ordinal'>
</column>
<column caption='State Name' datatype='string' name='[StateName]' role='dimension' semantic-role='[State].[Name]' type='nominal'>
</column>
<column caption='Predicted Payout' datatype='real' name='[Sum Total Paid (copy)]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' args <- data.frame(MonthlyDate = as.Date(.arg1, "%Y-%m-%d"), StateName = .arg2, LossCode = .arg3, stringsAsFactors = FALSE) ans <- rep(NA_real_, nrow(args)) for (i in seq_along(ans)) { val <- subset(devAggr, MonthlyDate == args[i, 1L] & StateName == args[i, 2L] & LossCode == args[i, 3L], "SumPredicted", drop = TRUE) if (length(val) == 1L) ans[i] <- val } ans ', MAX([MonthlyDate]), MAX([StateName]), MAX([LossCode]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column-instance column='[StateName]' derivation='Attribute' name='[attr:StateName:nk]' pivot='key' type='nominal' />
<column-instance column='[MonthlyDate]' derivation='MY' name='[my:MonthlyDate:ok]' pivot='key' type='ordinal' />
<column-instance column='[LossCode]' derivation='None' name='[none:LossCode:nk]' pivot='key' type='nominal' />
<column-instance column='[StateName]' derivation='None' name='[none:StateName:nk]' pivot='key' type='nominal' />
<column-instance column='[Mean Deviance (copy)]' derivation='User' name='[usr:Mean Deviance (copy):qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
<column-instance column='[Sum Total Paid (copy)]' derivation='User' name='[usr:Sum Total Paid (copy):qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
</datasource-dependencies>
<filter class='categorical' column='[textscan.41885.651995601853].[Action (State Name)]'>
<groupfilter function='member' level='[StateName]' member='"Wisconsin"' user:ui-action-filter='[Action1]' user:ui-domain='relevant' user:ui-enumeration='inclusive' user:ui-marker='enumerate' />
</filter>
<filter class='categorical' column='[textscan.41885.651995601853].[my:MonthlyDate:ok]'>
<groupfilter function='member' level='[my:MonthlyDate:ok]' member='201005' user:ui-domain='relevant' user:ui-enumeration='inclusive' user:ui-marker='enumerate' />
</filter>
<filter class='categorical' column='[textscan.41885.651995601853].[none:StateName:nk]'>
<groupfilter function='level-members' level='[none:StateName:nk]' user:ui-enumeration='all' user:ui-marker='enumerate' />
</filter>
<slices>
<column>[textscan.41885.651995601853].[my:MonthlyDate:ok]</column>
<column>[textscan.41885.651995601853].[none:StateName:nk]</column>
<column>[textscan.41885.651995601853].[Action (State Name)]</column>
</slices>
<aggregation value='true' />
</view>
<style>
<style-rule element='axis'>
<format attr='title' class='0' field='[textscan.41885.651995601853].[usr:Sum Total Paid (copy):qk]' scope='rows' value='' />
<format attr='subtitle' class='0' field='[textscan.41885.651995601853].[usr:Sum Total Paid (copy):qk]' scope='rows' value='' />
<format attr='auto-subtitle' class='0' field='[textscan.41885.651995601853].[usr:Sum Total Paid (copy):qk]' scope='rows' value='false' />
<encoding attr='space' class='0' field='[textscan.41885.651995601853].[usr:Sum Total Paid (copy):qk]' field-type='quantitative' fold='true' scope='rows' synchronized='true' type='space' />
<format attr='title' class='0' field='[textscan.41885.651995601853].[usr:Mean Deviance (copy):qk]' scope='rows' value='Payment ($)' />
<format attr='subtitle' class='0' field='[textscan.41885.651995601853].[usr:Mean Deviance (copy):qk]' scope='rows' value='' />
<format attr='auto-subtitle' class='0' field='[textscan.41885.651995601853].[usr:Mean Deviance (copy):qk]' scope='rows' value='true' />
</style-rule>
<style-rule element='legend-title-text'>
<format attr='color' field='[textscan.41885.651995601853].[:Measure Names]' value='Total Payments'>
<formatted-text>
<run>Total Payments</run>
</formatted-text>
</format>
</style-rule>
</style>
<panes>
<pane>
<view>
<breakdown value='auto' />
</view>
<mark class='Line' />
<encodings>
<color column='[textscan.41885.651995601853].[:Measure Names]' />
<tooltip column='[textscan.41885.651995601853].[attr:StateName:nk]' />
</encodings>
</pane>
<pane id='1' y-axis-name='[textscan.41885.651995601853].[usr:Mean Deviance (copy):qk]'>
<view>
<breakdown value='auto' />
</view>
<mark class='Line' />
<encodings>
<color column='[textscan.41885.651995601853].[:Measure Names]' />
<tooltip column='[textscan.41885.651995601853].[attr:StateName:nk]' />
</encodings>
</pane>
<pane id='2' y-axis-name='[textscan.41885.651995601853].[usr:Sum Total Paid (copy):qk]'>
<view>
<breakdown value='auto' />
</view>
<mark class='Line' />
<encodings>
<color column='[textscan.41885.651995601853].[:Measure Names]' />
<tooltip column='[textscan.41885.651995601853].[attr:StateName:nk]' />
</encodings>
</pane>
</panes>
<rows>([textscan.41885.651995601853].[usr:Mean Deviance (copy):qk] + [textscan.41885.651995601853].[usr:Sum Total Paid (copy):qk])</rows>
<cols>[textscan.41885.651995601853].[none:LossCode:nk]</cols>
</table>
</worksheet>
<worksheet name='Variable Importance'>
<table>
<view>
<datasources>
<datasource caption='claims_coefficients' name='textscan.41886.636795671293' />
<datasource name='Parameters' />
<datasource caption='claims_metadata' name='textscan.41885.651995601853' />
</datasources>
<datasource-dependencies datasource='textscan.41885.651995601853'>
<column caption='H2O Initialization' datatype='boolean' name='[Calculation_1010902082708389]' role='measure' type='nominal'>
<calculation class='tableau' formula='SCRIPT_BOOL( ' myh2oCleaner <- function(pattern = "Last.value.") { keys <- h2o.ls(h2oHandle, pattern = pattern)$Key if (!is.null(keys)) h2o.rm(h2oHandle, keys) invisible(keys) } if (!exists("h2oHandle", globalenv())) { library(h2o) h2oHandle <- h2o.init(ip = .arg1, port = .arg2, max_mem_size = .arg3, nthreads = -1) } if (length(h2o.ls(h2oHandle, pattern = "claims.dat")) == 0L) { claims.dat <- h2o.importFile(h2oHandle, path = .arg4, key = "claims.dat") # Create indicator variables for claim payout percentages invisible( h2o.exec(claims.dat$ClaimPaidRatio <- claims.dat$TotalPaid / claims.dat$TotalClaim) ) payoutThreshold <- 1 invisible( h2o.exec(claims.dat$HighClaimPaid <- claims.dat$ClaimPaidRatio > payoutThreshold) ) # Create log transformed variables invisible( h2o.exec(claims.dat$logDeduct <- log(claims.dat$Deduct + 1)) ) invisible( h2o.exec(claims.dat$logFees <- log(claims.dat$Fees + 1)) ) invisible( h2o.exec(claims.dat$logTotalClaim <- log(claims.dat$TotalClaim + 1)) ) invisible( h2o.exec(claims.dat$NonNegPaid <- ifelse(claims.dat$TotalPaid < 0, 0, claims.dat$TotalPaid)) ) invisible( h2o.exec(claims.dat$logTotalPaid <- log(claims.dat$NonNegPaid + 1)) ) # Convert some numeric columns to factor for (j in c("CallCenterPostalCode")) claims.dat[[j]] <- as.factor(claims.dat[[j]]) myh2oCleaner() # Limit analysis to "Claim Complete" completed.claims.dat <- claims.dat[claims.dat$ResponseStatus == "Claim Complete", ] completed.claims.dat <- h2o.assign(completed.claims.dat, "completed.claims.dat") myh2oCleaner() xvars <- c("CallCenterPostalCode", "Demographics_AgeGroup", "Demographics_Gender", "FirstClaim", "INStype", "LossCode", "MonthlyDate", "SourceCode", "StateName", "logDeduct", "logFees", "logTotalClaim") tweedie.final <- h2o.glm(xvars, "NonNegPaid", completed.claims.dat, key = "tweedie.final", family = "tweedie", tweedie.p = 1.75, alpha = 0, lambda = 12, variable_importances = TRUE) tweedie.vi <- sort(abs(tweedie.final@model$normalized_coefficients), decreasing = TRUE) tweedie.vi <- tweedie.vi[names(tweedie.vi) != "Intercept"] myh2oCleaner(pattern = "GLM2Job__") completed.claims.dat$predicted <- h2o.predict(tweedie.final, completed.claims.dat) myh2oCleaner(pattern = "GLM2Predict_") library(tweedie) completed.claims.dat$deviance <- tweedie.dev(y = completed.claims.dat$NonNegPaid, mu = completed.claims.dat$predicted, power = 1.75) claims.tweediedev.dat <- completed.claims.dat[!is.na(completed.claims.dat$deviance), ] claims.tweediedev.dat <- h2o.assign(claims.tweediedev.dat, "claims.tweediedev.dat") myh2oCleaner() devAggrFun <- function(x) { cbind(N = nrow(x), SumTotalPaid = sum(x$TotalPaid), SumPredicted = sum(x$predicted), MinDeviance = min(x$deviance), MeanDeviance = mean(x$deviance), MaxDeviance = max(x$deviance)) } h2o.addFunction(h2oHandle, devAggrFun) devAggr <- as.data.frame(h2o.ddply(claims.tweediedev.dat, c("MonthlyDate", "StateName", "LossCode"), devAggrFun)) devAggr[[1L]] <- as.Date(as.character(devAggr[[1L]]), "%m/%d/%Y") devAggr[[4L]] <- as.integer(devAggr[[4L]]) colnames(devAggr)[-(1:3)] <- c("N", "SumTotalPaid", "SumPredicted", "MinDeviance", "MeanDeviance", "MaxDeviance") print(head(devAggr, 10)) myh2oCleaner() } TRUE ', [Parameters].[Parameter 2], [Parameters].[Parameter 1], [Parameters].[Parameter 3], [Parameters].[Parameter 4])' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column-instance column='[Calculation_1010902082708389]' derivation='User' name='[usr:Calculation_1010902082708389:nk]' pivot='key' type='nominal'>
<table-calc ordering-type='Rows' />
</column-instance>
</datasource-dependencies>
<datasource-dependencies datasource='textscan.41886.636795671293'>
<column caption='Coefficient Name' datatype='string' name='[Coefficients]' role='dimension' type='nominal'>
</column>
<column caption='Normalized Value' datatype='real' name='[Number of Claims Filed (copy) 1]' role='measure' type='quantitative'>
<calculation class='tableau' formula='SCRIPT_REAL( ' tweedie.vi[.arg1] ', MAX([Coefficients]))' scope-isolation='false'>
<table-calc ordering-type='Rows' />
</calculation>
</column>
<column-instance column='[Coefficients]' derivation='None' name='[none:Coefficients:nk]' pivot='key' type='nominal' />
<column-instance column='[Number of Claims Filed (copy) 1]' derivation='User' name='[usr:Number of Claims Filed (copy) 1:qk]' pivot='key' type='quantitative'>
<table-calc ordering-type='Rows' />
</column-instance>
</datasource-dependencies>
<sort class='natural' column='[textscan.41886.636795671293].[none:Coefficients:nk]' direction='ASC' />
<aggregation value='true' />
</view>
<style>
</style>
<panes>
<pane>
<view>
<breakdown value='auto' />
</view>
<mark class='Automatic' />
<encodings>
<lod column='[textscan.41885.651995601853].[usr:Calculation_1010902082708389:nk]' />
</encodings>
</pane>
</panes>
<rows>[textscan.41886.636795671293].[usr:Number of Claims Filed (copy) 1:qk]</rows>
<cols>[textscan.41886.636795671293].[none:Coefficients:nk]</cols>
</table>
</worksheet>
</worksheets>
<dashboards>
<dashboard name='Predicted vs Actual'>
<style>
</style>
<size maxheight='800' maxwidth='1000' minheight='800' minwidth='1000' />
<zones>
<zone h='100000' id='2' type='layout-basic' w='100000' x='0' y='0'>
<zone h='100000' id='5' param='horz' type='layout-flow' w='100000' x='0' y='0'>
<zone h='100000' id='3' type='layout-basic' w='84000' x='0' y='0'>
<zone h='50000' id='16' name='Maximum Deviance' show-title='true' w='84000' x='0' y='0'>
</zone>
<zone h='50000' id='24' name='Payments by Loss Code' show-title='true' w='84000' x='0' y='50000'>
</zone>
</zone>
<zone fixed-size='160' h='100000' id='14' is-fixed='true' type='layout-basic' w='16000' x='84000' y='0'>
<zone h='100000' id='27' type='layout-basic' w='16000' x='84000' y='0'>
<zone h='58125' id='15' param='vert' type='layout-flow' w='16000' x='84000' y='0'>
<zone h='7375' id='17' name='Maximum Deviance' param='[textscan.41885.651995601853].[my:MonthlyDate:ok]' type='filter' w='16000' x='84000' y='0'>
</zone>
<zone fixed-size='133' h='17625' id='19' is-fixed='true' name='Maximum Deviance' pane-specification-id='0' param='[textscan.41885.651995601853].[usr:Number of Claims Filed (copy) 1 1:qk]' type='color' w='16000' x='84000' y='7375'>
</zone>
<zone h='6375' id='20' param='[Parameters].[Parameter 4]' type='paramctrl' w='16000' x='84000' y='25000'>
</zone>
<zone h='6375' id='21' param='[Parameters].[Parameter 2]' type='paramctrl' w='16000' x='84000' y='31375'>
</zone>
<zone h='6375' id='22' param='[Parameters].[Parameter 3]' type='paramctrl' w='16000' x='84000' y='37750'>
</zone>
<zone fixed-size='76' h='10500' id='23' is-fixed='true' param='[Parameters].[Parameter 1]' type='paramctrl' w='16000' x='84000' y='44125'>
</zone>
</zone>
<zone h='41875' id='28' param='vert' type='layout-flow' w='16000' x='84000' y='58125'>
<zone h='8250' id='26' name='Payments by Loss Code' pane-specification-id='1' param='[textscan.41885.651995601853].[:Measure Names]' type='color' w='16000' x='84000' y='58125'>
</zone>
</zone>
</zone>
</zone>
</zone>
</zone>
</zones>
</dashboard>
</dashboards>
<windows>
<window class='schema' source-height='-1' />
<window auto-hidden='0' class='dashboard' maximized='1' name='Predicted vs Actual'>
<zones>
<zone name='Maximum Deviance'>
<viewpoint>
<zoom type='entire-view' />
</viewpoint>
</zone>
<zone name='Payments by Loss Code'>
<viewpoint>
<zoom type='entire-view' />
</viewpoint>
</zone>
<zone mode='dropdown' name='Maximum Deviance' param='[textscan.41885.651995601853].[my:MonthlyDate:ok]' show-all='false' type='filter' />
<zone name='Maximum Deviance' pane-specification-id='0' param='[textscan.41885.651995601853].[usr:Number of Claims Filed (copy) 1 1:qk]' type='color' />
<zone mode='type_in' name='' param='[Parameters].[Parameter 4]' type='paramctrl' />
<zone mode='type_in' name='' param='[Parameters].[Parameter 2]' type='paramctrl' />
<zone mode='type_in' name='' param='[Parameters].[Parameter 3]' type='paramctrl' />
<zone mode='type_in' name='' param='[Parameters].[Parameter 1]' type='paramctrl' />
<zone name='Payments by Loss Code' pane-specification-id='1' param='[textscan.41885.651995601853].[:Measure Names]' type='color' />
</zones>
</window>
<window auto-hidden='0' class='worksheet' maximized='0' name='Variable Importance'>
<cards>
<edge name='left'>
<strip size='160'>
<card type='pages' />
<card type='filters' />
<card type='marks' />
</strip>
</edge>
<edge name='top'>
<strip size='31'>
<card type='columns' />
</strip>
<strip size='31'>
<card type='rows' />
</strip>
</edge>
<edge name='right'>
<strip size='160'>
<card mode='type_in' param='[Parameters].[Parameter 4]' type='parameter' />
<card mode='type_in' param='[Parameters].[Parameter 2]' type='parameter' />
<card mode='type_in' param='[Parameters].[Parameter 3]' type='parameter' />
<card mode='type_in' param='[Parameters].[Parameter 1]' type='parameter' />
</strip>
</edge>
</cards>
<highlight>
<color-one-way>
<field>[textscan.41885.651995601853].[usr:Calculation_1010902082708389:nk]</field>
<field>[textscan.41886.636795671293].[none:Coefficients:nk]</field>
</color-one-way>
</highlight>
</window>
<window class='hidden-worksheet' name='Maximum Deviance'>
<cards>
<edge name='left'>
<strip size='160'>
<card type='pages' />
<card type='filters' />
<card type='marks' />
<card pane-specification-id='0' param='[textscan.41885.651995601853].[usr:Number of Claims Filed (copy) 1 1:qk]' type='color' />
</strip>
</edge>
<edge name='top'>
<strip size='31'>
<card type='columns' />
</strip>
<strip size='31'>
<card type='rows' />
</strip>
</edge>
<edge name='right'>
<strip size='160'>
<card mode='dropdown' param='[textscan.41885.651995601853].[my:MonthlyDate:ok]' show-all='false' type='filter' />
<card mode='dropdown' param='[textscan.41885.651995601853].[none:LossCode:nk]' show-all='false' type='filter' />
<card mode='type_in' param='[Parameters].[Parameter 4]' type='parameter' />
<card mode='type_in' param='[Parameters].[Parameter 2]' type='parameter' />
<card mode='type_in' param='[Parameters].[Parameter 3]' type='parameter' />
<card mode='type_in' param='[Parameters].[Parameter 1]' type='parameter' />
</strip>
</edge>
</cards>
<viewpoint>
</viewpoint>
<highlight>
<color-one-way>
<field>[textscan.41885.651995601853].[attr:LossCode:nk]</field>
<field>[textscan.41885.651995601853].[my:MonthlyDate:ok]</field>
<field>[textscan.41885.651995601853].[none:LossCode:nk]</field>
<field>[textscan.41885.651995601853].[none:Region:nk]</field>
<field>[textscan.41885.651995601853].[usr:Calculation_1010902082708389:nk]</field>
<field>[textscan.41885.651995601853].[yr:MonthlyDate:ok]</field>
</color-one-way>
</highlight>
<selection-collection>
</selection-collection>
</window>
<window class='hidden-worksheet' name='Payments by Loss Code'>
<cards>
<edge name='left'>
<strip size='160'>
<card type='pages' />
<card type='filters' />
<card type='marks' />
</strip>
</edge>
<edge name='top'>
<strip size='31'>
<card type='columns' />
</strip>
<strip size='31'>
<card type='rows' />
</strip>
</edge>
<edge name='right'>
<strip size='160'>
<card mode='dropdown' param='[textscan.41885.651995601853].[my:MonthlyDate:ok]' show-all='false' type='filter' />
<card mode='dropdown' param='[textscan.41885.651995601853].[none:StateName:nk]' show-all='false' type='filter' />
<card pane-specification-id='1' param='[textscan.41885.651995601853].[:Measure Names]' type='color' />
</strip>
</edge>
</cards>
<viewpoint>
</viewpoint>
<highlight>
<color-one-way>
<field>[textscan.41885.651995601853].[my:MonthlyDate:ok]</field>
<field>[textscan.41885.651995601853].[none:LossCode:nk]</field>
<field>[textscan.41885.651995601853].[none:StateName:nk]</field>
</color-one-way>
</highlight>
<selection-collection>
</selection-collection>
</window>
</windows>
<thumbnails>
<thumbnail height='192' name='Predicted vs Actual' width='192'>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