In xqnwang/M5comp: Data from the M5 Competition

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

M5comp

M5 Competition Data

Installation

You can install the package M5comp with:

devtools::install_github("xqnwang/M5comp")

Aggregation

Number of M5 series per aggregation level.

| Level id | Aggregation level | Number of series | |:--------:|:---------------------------------------------------------------------|-----------------:| | 1 | Unit sales of all products, aggregated for all stores/states | 1 | | 2 | Unit sales of all products, aggregated for each State | 3 | | 3 | Unit sales of all products, aggregated for each store | 10 | | 4 | Unit sales of all products, aggregated for each category | 3 | | 5 | Unit sales of all products, aggregated for each department | 7 | | 6 | Unit sales of all products, aggregated for each State and category | 9 | | 7 | Unit sales of all products, aggregated for each State and department | 21 | | 8 | Unit sales of all products, aggregated for each store and category | 30 | | 9 | Unit sales of all products, aggregated for each store and department | 70 | | 10 | Unit sales of product x, aggregated for all stores/states | 3049 | | 11 | Unit sales of product x, aggregated for each State | 9147 | | 12 | Unit sales of product x, aggregated for each store | 30490 | | Total| - | 42840 |

The following table shows more simply and clearly how the aggregated time series are generated. For example, series of level 11 are aggregated by series with the same item.id and state.id.

| Level id | id | item.id | dept.id | cat.id | store.id | state.id | |:--------:|:--------:|:--------:|:--------:|:--------:|:--------:|:--------:| |(example) | (HOBBIES_1_001_CA_1) | (HOBBIES_1_001) | (HOBBIES_1) | (HOBBIES) | (CA_1) | (CA) | |1| | | | | | | |2| | | | | |$\checkmark$| |3| | | | |$\checkmark$| | |4| | | |$\checkmark$| | | |5| | |$\checkmark$| | | | |6| | | |$\checkmark$| |$\checkmark$| |7| | |$\checkmark$ | | |$\checkmark$ | |8| | | |$\checkmark$ |$\checkmark$ | | |9| | |$\checkmark$| |$\checkmark$| | |10| |$\checkmark$| | | | | |11| |$\checkmark$| | | |$\checkmark$| |12|$\checkmark$ | | | | | |

Example

The package M5comp contains two datasets: M5 and calendar.

M5

The M5 is a list of $42840$ series, including $30490$ bottom-level time series and $12350$ aggregated time series. Each element of this list is also a list with some components, like the time series of the training periods and the level id of the series. The components of the bottom-level series and the aggregated series are partially different.

# extract bottom-level series (level 12)
library(M5comp)
data(M5)
M5_bottom <- Filter(function(l) l$level == 12, M5)
# the structure of the bottom-level series
str(M5_bottom[[1]])

# extract extract level 9 series
M5_l9 <- Filter(function(l) l$level == 9, M5)
# the structure of the aggregated series
str(M5_l9[[1]])
# which levels the series is aggregated on
M5_l9[[1]]$agg.by

calendar

The calendar data contains information about the dates the products are sold. The calendar object is a dataframe with some components, such as date, weekday and event_name_1.

# the structure of the calendar data
data(calendar)
str(calendar)

# plot one of the series with date
library(ggplot2)
library(xts)
library(magrittr)
xts(M5[[1]]$x, order.by = calendar$date[1:M5[[1]]$n]) %>% 
   autoplot() + xlab("Time") + ylab("Sales")