Sync'd up the code in the various workbooks and making sure it all runs

.gitignore

@@ -1,4 +1,5 @@
 .Rproj.user/
+.vscode/
 
 *_cache/*
 *_files/*
@@ -12,3 +13,5 @@
 *.log
 
 temp*.R
+
+

Dockerfile

@@ -50,6 +50,7 @@ RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime \
     fs \
     furrr \
     loo \
+    markdown \
     modeltime \
     posterior \
     projpred \
@@ -70,8 +71,6 @@ RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime \
     timetk
 
 
-COPY build/logging.conf /etc/rstudio/
-
 WORKDIR /tmp
 
 COPY build/docker_install_sys_rpkgs.R /tmp

Makefile

@@ -58,6 +58,10 @@ construct_longsynth_onehier_pnbd_models.html: construct_longsynth_fixed_pnbd_mod
 construct_onlineretail_onehier_pnbd_models.html: construct_onlineretail_fixed_pnbd_models.html
 construct_cdnow_onehier_pnbd_models.html: construct_cdnow_fixed_pnbd_models.html
 
+construct_shortsynth_twohier_pnbd_models.html: construct_shortsynth_onehier_pnbd_models.html
+construct_longsynth_twohier_pnbd_models.html: construct_longsynth_onehier_pnbd_models.html
+construct_onlineretail_twohier_pnbd_models.html: construct_onlineretail_onehier_pnbd_models.html
+construct_cdnow_twohier_pnbd_models.html: construct_cdnow_onehier_pnbd_models.html
 
 
 mrproper: clean-cache clean-data clean-html clean-precompute clean-models

construct_cdnow_fixed_pnbd_models.qmd

@@ -85,7 +85,7 @@ try to model real-life data.
 
 ```{r set_start_end_dates}
 use_fit_start_date <- as.Date("1997-01-01")
-use_fit_end_date   <- as.Date("1997-12-31")
+use_fit_end_date   <- as.Date("1998-01-01")
 
 use_valid_start_date <- as.Date("1998-01-01")
 use_valid_end_date   <- as.Date("1998-07-01")

construct_cdnow_onehier_pnbd_models.qmd

@@ -83,7 +83,7 @@ try to model real-life data.
 
 ```{r set_start_end_dates}
 use_fit_start_date <- as.Date("1997-01-01")
-use_fit_end_date   <- as.Date("1997-12-31")
+use_fit_end_date   <- as.Date("1998-01-01")
 
 use_valid_start_date <- as.Date("1998-01-01")
 use_valid_end_date   <- as.Date("1999-01-10")

construct_longsynth_fixed_pnbd_models.qmd

@@ -82,15 +82,6 @@ synthetic data with the longer timeframe.
 We start by modelling the P/NBD model using our synthetic datasets before we
 try to model real-life data.
 
-```{r set_start_end_dates}
-use_fit_start_date <- as.Date("2010-01-01")
-use_fit_end_date   <- as.Date("2022-01-01")
-
-use_valid_start_date <- as.Date("2022-01-01")
-use_valid_end_date   <- as.Date("2023-01-01")
-```
-
-
 
 ## Load Long Time-frame Synthetic Data
 
@@ -111,9 +102,9 @@ customer_transactions_tbl |> glimpse()
 ```
 
 
-## Load Derived Datasets
+## Load Derived Data
 
-```{r write_data_disk}
+```{r load_modelling_data}
 #| echo: TRUE
 
 id_1000  <- read_rds("data/longsynth_id_1000.rds")
@@ -128,15 +119,48 @@ customer_summarystats_tbl <- read_rds("data/longsynth_customer_summarystats_tbl.
 
 obs_fitdata_tbl   <- read_rds("data/longsynth_obs_fitdata_tbl.rds")
 obs_validdata_tbl <- read_rds("data/longsynth_obs_validdata_tbl.rds")
+
+customer_subset_id <- id_1000
 ```
 
 
+## Load Subset Data
 
-# Fit First P/NBD Model
+We also want to construct our data subsets for the purposes of speeding up our
+valuations.
+
+```{r construct_customer_subset_data}
+#| echo: TRUE
+
+customer_fit_subset_tbl <- obs_fitdata_tbl |>
+  filter(customer_id %in% customer_subset_id)
+
+customer_fit_subset_tbl |> glimpse()
+
+
+customer_valid_subset_tbl <- obs_validdata_tbl |>
+  filter(customer_id %in% customer_subset_id)
+
+customer_valid_subset_tbl |> glimpse()
+```
+
+We now use these datasets to set the start and end dates for our various
+validation methods.
+
+
+```{r set_start_end_dates}
+dates_lst <- read_rds("data/longsynth_simulation_dates.rds")
+
+use_fit_start_date <- dates_lst$use_fit_start_date
+use_fit_end_date   <- dates_lst$use_fit_end_date
+
+use_valid_start_date <- dates_lst$use_valid_start_date
+use_valid_end_date   <- dates_lst$use_valid_end_date
+```
 
-We need to set up a number of workbook parameters, such as some directories
-for the Stan model.
 
+Finally, we need to set our directories where we save our Stan code and the
+model outputs.
 
 ```{r setup_workbook_parameters}
 #| echo: TRUE
@@ -147,6 +171,9 @@ stan_codedir  <-   "stan_code"
 
 
 
+# Fit First P/NBD Model
+
+
 ## Compile and Fit Stan Model
 
 We now compile this model using `CmdStanR`.
@@ -270,10 +297,12 @@ run the simulations and combine the datasets.
 ```{r run_pnbd_long_fixed1_assessment}
 #| echo: TRUE
 
+pnbd_stanfit <- pnbd_long_fixed1_stanfit |>
+  recover_types(customer_fit_stats_tbl)
+
 pnbd_long_fixed1_assess_data_lst <- run_model_assessment(
-  model_stanfit       = pnbd_long_fixed1_stanfit,
-  insample_tbl        = customer_fit_stats_tbl,
-  outsample_tbl       = customer_valid_stats_tbl,
+  model_stanfit       = pnbd_stanfit,
+  insample_tbl        = customer_fit_subset_tbl,
   fit_label           = "pnbd_long_fixed1",
   fit_end_dttm        = use_fit_end_date     |> as.POSIXct(),
   valid_start_dttm    = use_valid_start_date |> as.POSIXct(),
@@ -446,10 +475,12 @@ run the simulations and combine the datasets.
 ```{r run_pnbd_long_fixed2_assessment}
 #| echo: TRUE
 
+pnbd_stanfit <- pnbd_long_fixed2_stanfit |>
+  recover_types(customer_fit_stats_tbl)
+
 pnbd_long_fixed2_assess_data_lst <- run_model_assessment(
-  model_stanfit       = pnbd_long_fixed2_stanfit,
-  insample_tbl        = customer_fit_stats_tbl,
-  outsample_tbl       = customer_valid_stats_tbl,
+  model_stanfit       = pnbd_stanfit,
+  insample_tbl        = customer_fit_subset_tbl,
   fit_label           = "pnbd_long_fixed2",
   fit_end_dttm        = use_fit_end_date     |> as.POSIXct(),
   valid_start_dttm    = use_valid_start_date |> as.POSIXct(),
@@ -619,10 +650,12 @@ run the simulations and combine the datasets.
 ```{r run_pnbd_long_fixed3_assessment}
 #| echo: TRUE
 
+pnbd_stanfit <- pnbd_long_fixed3_stanfit |>
+  recover_types(customer_fit_stats_tbl)
+
 pnbd_long_fixed3_assess_data_lst <- run_model_assessment(
-  model_stanfit       = pnbd_long_fixed3_stanfit,
-  insample_tbl        = customer_fit_stats_tbl,
-  outsample_tbl       = customer_valid_stats_tbl,
+  model_stanfit       = pnbd_stanfit,
+  insample_tbl        = customer_fit_subset_tbl,
   fit_label           = "pnbd_long_fixed3",
   fit_end_dttm        = use_fit_end_date     |> as.POSIXct(),
   valid_start_dttm    = use_valid_start_date |> as.POSIXct(),