Anomalize a grouped tsibble #44
Description
I am using anomalize with a tsibble object. Since the data is grouped using the function index_by from tsibble anomalize() cannot work. This is due to an unsupported indexClass of type yearmonth.
I bring in data that is daily with no gaps
Here is my code:
> # Lib Load ####
> install.load::install_load(
+ "tidyquant"
+ , "fable"
+ , "fabletools"
+ , "feasts"
+ , "tsibble"
+ , "timetk"
+ , "sweep"
+ , "anomalize"
+ , "xts"
+ # , "fpp"
+ # , "forecast"
+ , "lubridate"
+ , "dplyr"
+ , "urca"
+ # , "prophet"
+ , "ggplot2"
+ )
> # Get File ####
> fileToLoad <- file.choose(new = TRUE)
> arrivals <- read.csv(fileToLoad)
> View(arrivals)
> arrivals$Time <- mdy(arrivals$Time)
> # Coerce to tsibble ----
> df_tsbl <- arrivals %>%
+ as_tsibble(index = Time)
> df_tsbl
# A tsibble: 6,908 x 2 [1D]
Time DSCH_COUNT
<date> <int>
1 2001-01-01 22
2 2001-01-02 30
3 2001-01-03 43
4 2001-01-04 30
5 2001-01-05 38
6 2001-01-06 22
7 2001-01-07 29
8 2001-01-08 37
9 2001-01-09 33
10 2001-01-10 52
# ... with 6,898 more rows
> interval(df_tsbl)
1D
> count_gaps(df_tsbl)
# A tibble: 0 x 3
# ... with 3 variables: .from <date>, .to <date>, .n <int>
> # Make Monthly ----
> df_monthly_tsbl <- df_tsbl %>%
+ index_by(Year_Month = ~ yearmonth(.)) %>%
+ summarise(Count = sum(DSCH_COUNT, na.rm = TRUE))
> df_monthly_tsbl
# A tsibble: 227 x 2 [1M]
Year_Month Count
<mth> <int>
1 2001 Jan 1067
2 2001 Feb 919
3 2001 Mar 1024
4 2001 Apr 1010
5 2001 May 1056
6 2001 Jun 995
7 2001 Jul 1002
8 2001 Aug 1076
9 2001 Sep 982
10 2001 Oct 971
# ... with 217 more rows
> # Anomalize ----
> df_monthly_tsbl %>%
+ time_decompose(Count, method = "twitter") %>%
+ anomalize(remainder, method = "gesd") %>%
+ clean_anomalies() %>%
+ time_recompose()
Converting from tbl_ts to tbl_time.
Auto-index message: index = Year_Month
Error in index.xts(x) : unsupported ‘indexClass’ indexing type: yearmonth
Session info:
> sessionInfo()
R version 3.5.3 (2019-03-11)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18362)
Matrix products: default
locale:
[1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] ggplot2_3.2.1 urca_1.3-0 dplyr_0.8.3
[4] anomalize_0.2.0 sweep_0.2.2 timetk_0.1.2
[7] tsibble_0.8.5 feasts_0.1.1 fable_0.1.1
[10] fabletools_0.1.1 tidyquant_0.5.9 quantmod_0.4-15
[13] TTR_0.23-6 PerformanceAnalytics_1.5.3 xts_0.11-2
[16] zoo_1.8-6 lubridate_1.7.4
loaded via a namespace (and not attached):
[1] tidyselect_0.2.5 purrr_0.3.3 lattice_0.20-38 colorspace_1.4-1
[5] vctrs_0.2.1 generics_0.0.2 utf8_1.1.4 rlang_0.4.2
[9] pillar_1.4.3 tibbletime_0.1.3 glue_1.3.1 withr_2.1.2
[13] lifecycle_0.1.0 stringr_1.4.0 Quandl_2.10.0 munsell_0.5.0
[17] anytime_0.3.6 gtable_0.3.0 labeling_0.3 curl_4.3
[21] fansi_0.4.1 broom_0.5.3 Rcpp_1.0.3 backports_1.1.5
[25] scales_1.1.0 install.load_1.2.1 jsonlite_1.6 farver_2.0.2
[29] gridExtra_2.3 digest_0.6.23 packrat_0.5.0 stringi_1.4.3
[33] grid_3.5.3 quadprog_1.5-8 cli_2.0.1 tools_3.5.3
[37] magrittr_1.5 lazyeval_0.2.2 tibble_2.1.3 crayon_1.3.4
[41] tidyr_1.0.0 pkgconfig_2.0.3 zeallot_0.1.0 assertthat_0.2.1
[45] httr_1.4.1 rstudioapi_0.10 R6_2.4.1 nlme_3.1-137
[49] compiler_3.5.3