healthyverse_tsa
Time Series Analysis, Modeling and Forecasting of the Healthyverse
Packages Steven P. Sanderson II, MPH - Date: 2025-12-05
Introduction
This analysis follows a Nested Modeltime Workflow from modeltime
along with using the NNS package. I use this to monitor the
downloads of all of my packages:
Get Data
glimpse(downloads_tbl)
Rows: 162,076
Columns: 11
$ date <date> 2020-11-23, 2020-11-23, 2020-11-23, 2020-11-23, 2020-11-23,…
$ time <Period> 15H 36M 55S, 11H 26M 39S, 23H 34M 44S, 18H 39M 32S, 9H 0M…
$ date_time <dttm> 2020-11-23 15:36:55, 2020-11-23 11:26:39, 2020-11-23 23:34:…
$ size <int> 4858294, 4858294, 4858301, 4858295, 361, 4863722, 4864794, 4…
$ r_version <chr> NA, "4.0.3", "3.5.3", "3.5.2", NA, NA, NA, NA, NA, NA, NA, N…
$ r_arch <chr> NA, "x86_64", "x86_64", "x86_64", NA, NA, NA, NA, NA, NA, NA…
$ r_os <chr> NA, "mingw32", "mingw32", "linux-gnu", NA, NA, NA, NA, NA, N…
$ package <chr> "healthyR.data", "healthyR.data", "healthyR.data", "healthyR…
$ version <chr> "1.0.0", "1.0.0", "1.0.0", "1.0.0", "1.0.0", "1.0.0", "1.0.0…
$ country <chr> "US", "US", "US", "GB", "US", "US", "DE", "HK", "JP", "US", …
$ ip_id <int> 2069, 2804, 78827, 27595, 90474, 90474, 42435, 74, 7655, 638…
The last day in the data set is 2025-12-03 23:41:55, the file was birthed on: 2022-07-02 23:58:17.511888, and at report knit time is 2.999573^{4} hours old. Happy analyzing!
Now that we have our data lets take a look at it using the skimr
package.
skim(downloads_tbl)
| Name | downloads_tbl |
| Number of rows | 162076 |
| Number of columns | 11 |
| _______________________ | |
| Column type frequency: | |
| character | 6 |
| Date | 1 |
| numeric | 2 |
| POSIXct | 1 |
| Timespan | 1 |
| ________________________ | |
| Group variables | None |
Data summary
Variable type: character
| skim_variable | n_missing | complete_rate | min | max | empty | n_unique | whitespace |
|---|---|---|---|---|---|---|---|
| r_version | 118839 | 0.27 | 5 | 7 | 0 | 50 | 0 |
| r_arch | 118839 | 0.27 | 1 | 7 | 0 | 6 | 0 |
| r_os | 118839 | 0.27 | 7 | 19 | 0 | 24 | 0 |
| package | 0 | 1.00 | 7 | 13 | 0 | 8 | 0 |
| version | 0 | 1.00 | 5 | 17 | 0 | 62 | 0 |
| country | 15202 | 0.91 | 2 | 2 | 0 | 166 | 0 |
Variable type: Date
| skim_variable | n_missing | complete_rate | min | max | median | n_unique |
|---|---|---|---|---|---|---|
| date | 0 | 1 | 2020-11-23 | 2025-12-03 | 2023-11-02 | 1830 |
Variable type: numeric
| skim_variable | n_missing | complete_rate | mean | sd | p0 | p25 | p50 | p75 | p100 | hist |
|---|---|---|---|---|---|---|---|---|---|---|
| size | 0 | 1 | 1124523.34 | 1488471.59 | 355 | 27381 | 310246.5 | 2353996 | 5677952 | ▇▁▂▁▁ |
| ip_id | 0 | 1 | 11333.74 | 21980.47 | 1 | 235 | 2889.0 | 11961 | 299146 | ▇▁▁▁▁ |
Variable type: POSIXct
| skim_variable | n_missing | complete_rate | min | max | median | n_unique |
|---|---|---|---|---|---|---|
| date_time | 0 | 1 | 2020-11-23 09:00:41 | 2025-12-03 23:41:55 | 2023-11-02 16:25:32 | 102264 |
Variable type: Timespan
| skim_variable | n_missing | complete_rate | min | max | median | n_unique |
|---|---|---|---|---|---|---|
| time | 0 | 1 | 0 | 59 | 48.5 | 60 |
We can see that the following columns are missing a lot of data and for
us are most likely not useful anyways, so we will drop them
c(r_version, r_arch, r_os)
Plots
Now lets take a look at a time-series plot of the total daily downloads by package. We will use a log scale and place a vertical line at each version release for each package.
[[1]]
[[2]]
[[3]]
[[4]]
[[5]]
[[6]]
[[7]]
[[8]]
Now lets take a look at some time series decomposition graphs.
[[1]]
[[2]]
[[3]]
[[4]]
[[5]]
[[6]]
[[7]]
[[8]]
[[1]]
[[2]]
[[3]]
[[4]]
[[5]]
[[6]]
[[7]]
[[8]]
Seasonal Diagnostics:
[[1]]
[[2]]
[[3]]
[[4]]
[[5]]
[[6]]
[[7]]
[[8]]
ACF and PACF Diagnostics:
[[1]]
[[2]]
[[3]]
[[4]]
[[5]]
[[6]]
[[7]]
[[8]]
Feature Engineering
Now that we have our basic data and a shot of what it looks like, let’s
add some features to our data which can be very helpful in modeling.
Lets start by making a tibble that is aggregated by the day and
package, as we are going to be interested in forecasting the next 4
weeks or 28 days for each package. First lets get our base data.
Call:
stats::lm(formula = .formula, data = df)
Residuals:
Min 1Q Median 3Q Max
-146.36 -36.48 -11.22 27.08 819.45
Coefficients:
Estimate Std. Error
(Intercept) -1.822e+02 6.072e+01
date 1.117e-02 3.219e-03
lag(value, 1) 1.096e-01 2.323e-02
lag(value, 7) 9.047e-02 2.395e-02
lag(value, 14) 7.620e-02 2.391e-02
lag(value, 21) 7.997e-02 2.397e-02
lag(value, 28) 6.843e-02 2.393e-02
lag(value, 35) 5.573e-02 2.398e-02
lag(value, 42) 6.137e-02 2.409e-02
lag(value, 49) 6.288e-02 2.398e-02
month(date, label = TRUE).L -1.062e+01 5.040e+00
month(date, label = TRUE).Q 4.561e-01 4.974e+00
month(date, label = TRUE).C -1.588e+01 5.002e+00
month(date, label = TRUE)^4 -6.124e+00 4.961e+00
month(date, label = TRUE)^5 -6.705e+00 4.910e+00
month(date, label = TRUE)^6 1.253e+00 4.916e+00
month(date, label = TRUE)^7 -4.497e+00 4.849e+00
month(date, label = TRUE)^8 -4.016e+00 4.821e+00
month(date, label = TRUE)^9 2.766e+00 4.834e+00
month(date, label = TRUE)^10 9.302e-01 4.850e+00
month(date, label = TRUE)^11 -4.077e+00 4.837e+00
fourier_vec(date, type = "sin", K = 1, period = 7) -1.126e+01 2.219e+00
fourier_vec(date, type = "cos", K = 1, period = 7) 7.347e+00 2.298e+00
t value Pr(>|t|)
(Intercept) -3.000 0.002734 **
date 3.469 0.000534 ***
lag(value, 1) 4.720 2.54e-06 ***
lag(value, 7) 3.777 0.000164 ***
lag(value, 14) 3.187 0.001462 **
lag(value, 21) 3.335 0.000869 ***
lag(value, 28) 2.860 0.004285 **
lag(value, 35) 2.324 0.020237 *
lag(value, 42) 2.548 0.010931 *
lag(value, 49) 2.622 0.008820 **
month(date, label = TRUE).L -2.106 0.035311 *
month(date, label = TRUE).Q 0.092 0.926953
month(date, label = TRUE).C -3.175 0.001527 **
month(date, label = TRUE)^4 -1.234 0.217263
month(date, label = TRUE)^5 -1.366 0.172260
month(date, label = TRUE)^6 0.255 0.798826
month(date, label = TRUE)^7 -0.927 0.353807
month(date, label = TRUE)^8 -0.833 0.404939
month(date, label = TRUE)^9 0.572 0.567222
month(date, label = TRUE)^10 0.192 0.847931
month(date, label = TRUE)^11 -0.843 0.399324
fourier_vec(date, type = "sin", K = 1, period = 7) -5.075 4.28e-07 ***
fourier_vec(date, type = "cos", K = 1, period = 7) 3.196 0.001416 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 59.33 on 1758 degrees of freedom
(49 observations deleted due to missingness)
Multiple R-squared: 0.2302, Adjusted R-squared: 0.2205
F-statistic: 23.89 on 22 and 1758 DF, p-value: < 2.2e-16
NNS Forecasting
This is something I have been wanting to try for a while. The NNS
package is a great package for forecasting time series data.
library(NNS)
data_list <- base_data |>
select(package, value) |>
group_split(package)
data_list |>
imap(
\(x, idx) {
obj <- x
x <- obj |> pull(value) |> tail(7*52)
train_set_size <- length(x) - 56
pkg <- obj |> pluck(1) |> unique()
# sf <- NNS.seas(x, modulo = 7, plot = FALSE)$periods
seas <- t(
sapply(
1:25,
function(i) c(
i,
sqrt(
mean((
NNS.ARMA(x,
h = 28,
training.set = train_set_size,
method = "lin",
seasonal.factor = i,
plot=FALSE
) - tail(x, 28)) ^ 2)))
)
)
colnames(seas) <- c("Period", "RMSE")
sf <- seas[which.min(seas[, 2]), 1]
cat(paste0("Package: ", pkg, "\n"))
NNS.ARMA.optim(
variable = x,
h = 28,
training.set = train_set_size,
#seasonal.factor = seq(12, 60, 7),
seasonal.factor = sf,
pred.int = 0.95,
plot = TRUE
)
title(
sub = paste0("\n",
"Package: ", pkg, " - NNS Optimization")
)
}
)
Package: healthyR
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 12 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 7.42247084181809"
[1] "BEST method = 'lin' PATH MEMBER = c( 12 )"
[1] "BEST lin OBJECTIVE FUNCTION = 7.42247084181809"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 12 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 7.77177300180908"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 12 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 7.77177300180908"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 12 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 6.80688289452525"
[1] "BEST method = 'both' PATH MEMBER = c( 12 )"
[1] "BEST both OBJECTIVE FUNCTION = 6.80688289452525"
Package: healthyR.ai
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 3 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 19.3101909072865"
[1] "BEST method = 'lin' PATH MEMBER = c( 3 )"
[1] "BEST lin OBJECTIVE FUNCTION = 19.3101909072865"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 3 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 5.88232472019315"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 3 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 5.88232472019315"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 3 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 18.54670480521"
[1] "BEST method = 'both' PATH MEMBER = c( 3 )"
[1] "BEST both OBJECTIVE FUNCTION = 18.54670480521"
Package: healthyR.data
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 5 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 25.6147734004602"
[1] "BEST method = 'lin' PATH MEMBER = c( 5 )"
[1] "BEST lin OBJECTIVE FUNCTION = 25.6147734004602"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 5 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 8.64632847859436"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 5 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 8.64632847859436"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 5 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 7.9302272671102"
[1] "BEST method = 'both' PATH MEMBER = c( 5 )"
[1] "BEST both OBJECTIVE FUNCTION = 7.9302272671102"
Package: healthyR.ts
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 1 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 110.832998065742"
[1] "BEST method = 'lin' PATH MEMBER = c( 1 )"
[1] "BEST lin OBJECTIVE FUNCTION = 110.832998065742"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 1 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 77.0627951140239"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 1 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 77.0627951140239"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 1 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 87.8727909158634"
[1] "BEST method = 'both' PATH MEMBER = c( 1 )"
[1] "BEST both OBJECTIVE FUNCTION = 87.8727909158634"
Package: healthyverse
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 12 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 6.02402780869286"
[1] "BEST method = 'lin' PATH MEMBER = c( 12 )"
[1] "BEST lin OBJECTIVE FUNCTION = 6.02402780869286"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 12 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 11.3824053370165"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 12 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 11.3824053370165"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 12 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 8.72458943634062"
[1] "BEST method = 'both' PATH MEMBER = c( 12 )"
[1] "BEST both OBJECTIVE FUNCTION = 8.72458943634062"
Package: RandomWalker
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 13 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 11.3718625767837"
[1] "BEST method = 'lin' PATH MEMBER = c( 13 )"
[1] "BEST lin OBJECTIVE FUNCTION = 11.3718625767837"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 13 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 6.12876584702821"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 13 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 6.12876584702821"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 13 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 6.41728184647416"
[1] "BEST method = 'both' PATH MEMBER = c( 13 )"
[1] "BEST both OBJECTIVE FUNCTION = 6.41728184647416"
Package: tidyAML
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 2 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 59.1862487005951"
[1] "BEST method = 'lin' PATH MEMBER = c( 2 )"
[1] "BEST lin OBJECTIVE FUNCTION = 59.1862487005951"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 2 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 8.05237103840849"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 2 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 8.05237103840849"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 2 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 11.9019338128146"
[1] "BEST method = 'both' PATH MEMBER = c( 2 )"
[1] "BEST both OBJECTIVE FUNCTION = 11.9019338128146"
Package: TidyDensity
[1] "CURRNET METHOD: lin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'lin' , seasonal.factor = c( 16 ) ...)"
[1] "CURRENT lin OBJECTIVE FUNCTION = 10.6500373310391"
[1] "BEST method = 'lin' PATH MEMBER = c( 16 )"
[1] "BEST lin OBJECTIVE FUNCTION = 10.6500373310391"
[1] "CURRNET METHOD: nonlin"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'nonlin' , seasonal.factor = c( 16 ) ...)"
[1] "CURRENT nonlin OBJECTIVE FUNCTION = 19.1226264417811"
[1] "BEST method = 'nonlin' PATH MEMBER = c( 16 )"
[1] "BEST nonlin OBJECTIVE FUNCTION = 19.1226264417811"
[1] "CURRNET METHOD: both"
[1] "COPY LATEST PARAMETERS DIRECTLY FOR NNS.ARMA() IF ERROR:"
[1] "NNS.ARMA(... method = 'both' , seasonal.factor = c( 16 ) ...)"
[1] "CURRENT both OBJECTIVE FUNCTION = 13.548463171259"
[1] "BEST method = 'both' PATH MEMBER = c( 16 )"
[1] "BEST both OBJECTIVE FUNCTION = 13.548463171259"
[[1]]
NULL
[[2]]
NULL
[[3]]
NULL
[[4]]
NULL
[[5]]
NULL
[[6]]
NULL
[[7]]
NULL
[[8]]
NULL
Pre-Processing
Now we are going to do some basic pre-processing.
data_padded_tbl <- base_data %>%
pad_by_time(
.date_var = date,
.pad_value = 0
)
# Get log interval and standardization parameters
log_params <- liv(data_padded_tbl$value, limit_lower = 0, offset = 1, silent = TRUE)
limit_lower <- log_params$limit_lower
limit_upper <- log_params$limit_upper
offset <- log_params$offset
data_liv_tbl <- data_padded_tbl %>%
# Get log interval transform
mutate(value_trans = liv(value, limit_lower = 0, offset = 1, silent = TRUE)$log_scaled)
# Get Standardization Params
std_params <- standard_vec(data_liv_tbl$value_trans, silent = TRUE)
std_mean <- std_params$mean
std_sd <- std_params$sd
data_transformed_tbl <- data_liv_tbl %>%
group_by(package) %>%
# get standardization
mutate(value_trans = standard_vec(value_trans, silent = TRUE)$standard_scaled) %>%
tk_augment_fourier(
.date_var = date,
.periods = c(7, 14, 30, 90, 180),
.K = 2
) %>%
tk_augment_timeseries_signature(
.date_var = date
) %>%
ungroup() %>%
select(-c(value, -year.iso))
Since this is panel data we can follow one of two different modeling strategies. We can search for a global model in the panel data or we can use nested forecasting finding the best model for each of the time series. Since we only have 5 panels, we will use nested forecasting.
To do this we will use the nest_timeseries and
split_nested_timeseries functions to create a nested tibble.
horizon <- 4*7
nested_data_tbl <- data_transformed_tbl %>%
# 0. Filter out column where package is NA
filter(!is.na(package)) %>%
# 1. Extending: We'll predict n days into the future.
extend_timeseries(
.id_var = package,
.date_var = date,
.length_future = horizon
) %>%
# 2. Nesting: We'll group by id, and create a future dataset
# that forecasts n days of extended data and
# an actual dataset that contains n*2 days
nest_timeseries(
.id_var = package,
.length_future = horizon
#.length_actual = horizon*2
) %>%
# 3. Splitting: We'll take the actual data and create splits
# for accuracy and confidence interval estimation of n das (test)
# and the rest is training data
split_nested_timeseries(
.length_test = horizon
)
nested_data_tbl
# A tibble: 8 × 4
package .actual_data .future_data .splits
<fct> <list> <list> <list>
1 healthyR.data <tibble [1,822 × 50]> <tibble [28 × 50]> <split [1794|28]>
2 healthyR <tibble [1,813 × 50]> <tibble [28 × 50]> <split [1785|28]>
3 healthyR.ts <tibble [1,758 × 50]> <tibble [28 × 50]> <split [1730|28]>
4 healthyverse <tibble [1,729 × 50]> <tibble [28 × 50]> <split [1701|28]>
5 healthyR.ai <tibble [1,555 × 50]> <tibble [28 × 50]> <split [1527|28]>
6 TidyDensity <tibble [1,406 × 50]> <tibble [28 × 50]> <split [1378|28]>
7 tidyAML <tibble [1,013 × 50]> <tibble [28 × 50]> <split [985|28]>
8 RandomWalker <tibble [436 × 50]> <tibble [28 × 50]> <split [408|28]>
Now it is time to make some recipes and models using the modeltime workflow.
Modeltime Workflow
Recipe Object
recipe_base <- recipe(
value_trans ~ .
, data = extract_nested_test_split(nested_data_tbl)
)
recipe_base
recipe_date <- recipe(
value_trans ~ date
, data = extract_nested_test_split(nested_data_tbl)
)
Models
# Models ------------------------------------------------------------------
# Auto ARIMA --------------------------------------------------------------
model_spec_arima_no_boost <- arima_reg() %>%
set_engine(engine = "auto_arima")
wflw_auto_arima <- workflow() %>%
add_recipe(recipe = recipe_date) %>%
add_model(model_spec_arima_no_boost)
# NNETAR ------------------------------------------------------------------
model_spec_nnetar <- nnetar_reg(
mode = "regression"
, seasonal_period = "auto"
) %>%
set_engine("nnetar")
wflw_nnetar <- workflow() %>%
add_recipe(recipe = recipe_base) %>%
add_model(model_spec_nnetar)
# TSLM --------------------------------------------------------------------
model_spec_lm <- linear_reg() %>%
set_engine("lm")
wflw_lm <- workflow() %>%
add_recipe(recipe = recipe_base) %>%
add_model(model_spec_lm)
# MARS --------------------------------------------------------------------
model_spec_mars <- mars(mode = "regression") %>%
set_engine("earth")
wflw_mars <- workflow() %>%
add_recipe(recipe = recipe_date) %>%
add_model(model_spec_mars)
Nested Modeltime Tables
nested_modeltime_tbl <- modeltime_nested_fit(
# Nested Data
nested_data = nested_data_tbl,
control = control_nested_fit(
verbose = TRUE,
allow_par = FALSE
),
# Add workflows
wflw_auto_arima,
wflw_lm,
wflw_mars,
wflw_nnetar
)
nested_modeltime_tbl <- nested_modeltime_tbl[!is.na(nested_modeltime_tbl$package),]
Model Accuracy
nested_modeltime_tbl %>%
extract_nested_test_accuracy() %>%
filter(!is.na(package)) %>%
knitr::kable()
| package | .model_id | .model_desc | .type | mae | mape | mase | smape | rmse | rsq |
|---|---|---|---|---|---|---|---|---|---|
| healthyR.data | 1 | ARIMA | Test | 0.9271055 | 142.47142 | 0.7873815 | 175.7471 | 1.0944280 | 0.0234789 |
| healthyR.data | 2 | LM | Test | 0.8092850 | 160.90353 | 0.6873177 | 133.3049 | 0.9357054 | 0.0259016 |
| healthyR.data | 3 | EARTH | Test | 0.8931487 | 108.38546 | 0.7585423 | 168.9094 | 1.0557934 | 0.0003864 |
| healthyR.data | 4 | NNAR | Test | 0.8374680 | 211.86866 | 0.7112532 | 121.9725 | 0.9515265 | 0.0050638 |
| healthyR | 1 | ARIMA | Test | 0.8512157 | 144.94551 | 0.8221898 | 160.4913 | 1.1334958 | 0.0084631 |
| healthyR | 2 | LM | Test | 0.7636841 | 449.29238 | 0.7376430 | 128.3083 | 0.9697839 | 0.0341594 |
| healthyR | 3 | EARTH | Test | 1.0624454 | 304.75338 | 1.0262167 | 178.5461 | 1.3259972 | 0.0267311 |
| healthyR | 4 | NNAR | Test | 0.8091714 | 455.17968 | 0.7815792 | 136.3280 | 1.0294248 | 0.0077425 |
| healthyR.ts | 1 | ARIMA | Test | 0.6965950 | 182.62838 | 0.6797742 | 145.9751 | 0.9199612 | 0.0252838 |
| healthyR.ts | 2 | LM | Test | 0.8742884 | 344.00198 | 0.8531767 | 162.4294 | 1.0533880 | 0.0040476 |
| healthyR.ts | 3 | EARTH | Test | 0.6421260 | 126.91347 | 0.6266204 | 136.4706 | 0.8835031 | 0.0124713 |
| healthyR.ts | 4 | NNAR | Test | 0.8701253 | 342.84741 | 0.8491141 | 157.8798 | 1.0384783 | 0.0137563 |
| healthyverse | 1 | ARIMA | Test | 0.9151274 | 91.74449 | 0.9319919 | 163.3505 | 1.1587324 | 0.0891618 |
| healthyverse | 2 | LM | Test | 0.9446937 | 158.51547 | 0.9621031 | 143.2456 | 1.0921942 | 0.0123855 |
| healthyverse | 3 | EARTH | Test | 0.7849667 | 142.19034 | 0.7994324 | 113.8378 | 0.9833319 | 0.0003917 |
| healthyverse | 4 | NNAR | Test | 0.9211702 | 154.49985 | 0.9381460 | 133.7985 | 1.0880846 | 0.0016442 |
| healthyR.ai | 1 | ARIMA | Test | 0.9848080 | 99.46817 | 1.0848422 | 190.3269 | 1.1748352 | 0.0026857 |
| healthyR.ai | 2 | LM | Test | 1.2125269 | 151.83698 | 1.3356923 | 165.6941 | 1.3996556 | 0.0322248 |
| healthyR.ai | 3 | EARTH | Test | 0.8128680 | 81.32210 | 0.8954370 | 113.5764 | 1.0382718 | 0.1798390 |
| healthyR.ai | 4 | NNAR | Test | 1.2164298 | 150.14213 | 1.3399916 | 168.5875 | 1.4070368 | 0.0003136 |
| TidyDensity | 1 | ARIMA | Test | 0.9605820 | 265.67435 | 0.6116639 | 143.5840 | 1.1426382 | 0.5243825 |
| TidyDensity | 2 | LM | Test | 0.8768535 | 118.80247 | 0.5583486 | 150.0834 | 1.0675961 | 0.0803710 |
| TidyDensity | 3 | EARTH | Test | 1.1284944 | 317.74176 | 0.7185844 | 128.9309 | 1.3901836 | 0.0645247 |
| TidyDensity | 4 | NNAR | Test | 0.9308894 | 168.92310 | 0.5927567 | 149.6938 | 1.0814238 | 0.0716393 |
| tidyAML | 1 | ARIMA | Test | 0.8779933 | 119.07906 | 1.0025434 | 184.4475 | 1.0370597 | 0.0158876 |
| tidyAML | 2 | LM | Test | 0.7195456 | 206.69579 | 0.8216187 | 124.2401 | 0.8708057 | 0.0818809 |
| tidyAML | 3 | EARTH | Test | 0.9622259 | 158.90260 | 1.0987252 | 181.7720 | 1.1179834 | 0.0041723 |
| tidyAML | 4 | NNAR | Test | 0.7217111 | 191.02704 | 0.8240914 | 123.7270 | 0.8663225 | 0.0681259 |
| RandomWalker | 1 | ARIMA | Test | 0.8532957 | 127.71923 | 0.7577963 | 166.6293 | 0.9169701 | 0.2001317 |
| RandomWalker | 2 | LM | Test | 0.8987630 | 146.99003 | 0.7981749 | 157.4173 | 1.0045599 | 0.0005473 |
| RandomWalker | 3 | EARTH | Test | 0.9890069 | 167.18689 | 0.8783189 | 163.0656 | 1.0424660 | 0.0639651 |
| RandomWalker | 4 | NNAR | Test | 1.0058047 | 183.52939 | 0.8932367 | 164.1921 | 1.1041968 | 0.0030022 |
Plot Models
nested_modeltime_tbl %>%
extract_nested_test_forecast() %>%
group_by(package) %>%
filter_by_time(.date_var = .index, .start_date = max(.index) - 60) %>%
ungroup() %>%
plot_modeltime_forecast(
.interactive = FALSE,
.conf_interval_show = FALSE,
.facet_scales = "free"
) +
theme_minimal() +
facet_wrap(~ package, nrow = 3) +
theme(legend.position = "bottom")
Best Model
best_nested_modeltime_tbl <- nested_modeltime_tbl %>%
modeltime_nested_select_best(
metric = "rmse",
minimize = TRUE,
filter_test_forecasts = TRUE
)
best_nested_modeltime_tbl %>%
extract_nested_best_model_report()
# Nested Modeltime Table
# A tibble: 8 × 10
package .model_id .model_desc .type mae mape mase smape rmse rsq
<fct> <int> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 healthyR.da… 2 LM Test 0.809 161. 0.687 133. 0.936 2.59e-2
2 healthyR 2 LM Test 0.764 449. 0.738 128. 0.970 3.42e-2
3 healthyR.ts 3 EARTH Test 0.642 127. 0.627 136. 0.884 1.25e-2
4 healthyverse 3 EARTH Test 0.785 142. 0.799 114. 0.983 3.92e-4
5 healthyR.ai 3 EARTH Test 0.813 81.3 0.895 114. 1.04 1.80e-1
6 TidyDensity 2 LM Test 0.877 119. 0.558 150. 1.07 8.04e-2
7 tidyAML 4 NNAR Test 0.722 191. 0.824 124. 0.866 6.81e-2
8 RandomWalker 1 ARIMA Test 0.853 128. 0.758 167. 0.917 2.00e-1
best_nested_modeltime_tbl %>%
extract_nested_test_forecast() %>%
#filter(!is.na(.model_id)) %>%
group_by(package) %>%
filter_by_time(.date_var = .index, .start_date = max(.index) - 60) %>%
ungroup() %>%
plot_modeltime_forecast(
.interactive = FALSE,
.conf_interval_alpha = 0.2,
.facet_scales = "free"
) +
facet_wrap(~ package, nrow = 3) +
theme_minimal() +
theme(legend.position = "bottom")
Refitting and Future Forecast
Now that we have the best models, we can make our future forecasts.
nested_modeltime_refit_tbl <- best_nested_modeltime_tbl %>%
modeltime_nested_refit(
control = control_nested_refit(verbose = TRUE)
)
nested_modeltime_refit_tbl
# Nested Modeltime Table
# A tibble: 8 × 5
package .actual_data .future_data .splits .modeltime_tables
<fct> <list> <list> <list> <list>
1 healthyR.data <tibble> <tibble> <split [1794|28]> <mdl_tm_t [1 × 5]>
2 healthyR <tibble> <tibble> <split [1785|28]> <mdl_tm_t [1 × 5]>
3 healthyR.ts <tibble> <tibble> <split [1730|28]> <mdl_tm_t [1 × 5]>
4 healthyverse <tibble> <tibble> <split [1701|28]> <mdl_tm_t [1 × 5]>
5 healthyR.ai <tibble> <tibble> <split [1527|28]> <mdl_tm_t [1 × 5]>
6 TidyDensity <tibble> <tibble> <split [1378|28]> <mdl_tm_t [1 × 5]>
7 tidyAML <tibble> <tibble> <split [985|28]> <mdl_tm_t [1 × 5]>
8 RandomWalker <tibble> <tibble> <split [408|28]> <mdl_tm_t [1 × 5]>
nested_modeltime_refit_tbl %>%
extract_nested_future_forecast() %>%
group_by(package) %>%
mutate(across(.value:.conf_hi, .fns = ~ standard_inv_vec(
x = .,
mean = std_mean,
sd = std_sd
)$standard_inverse_value)) %>%
mutate(across(.value:.conf_hi, .fns = ~ liiv(
x = .,
limit_lower = limit_lower,
limit_upper = limit_upper,
offset = offset
)$rescaled_v)) %>%
filter_by_time(.date_var = .index, .start_date = max(.index) - 60) %>%
ungroup() %>%
plot_modeltime_forecast(
.interactive = FALSE,
.conf_interval_alpha = 0.2,
.facet_scales = "free"
) +
facet_wrap(~ package, nrow = 3) +
theme_minimal() +
theme(legend.position = "bottom")
