CODEWITHSUNDEEP
rdplyrtime-seriesforecastingfacebook-prophet
Alex Ho
Alex Ho

Reputation: 438

Using Prophet package to forecast by groups and create plot

I am using Prophet package to forecasting in groups in a dataframe, and I want to create plots using the grouped dataframe.

I was following the answers in Using Prophet Package to Predict by Group in Dataframe in R. Are there easier ways to create the plots, compared to how I did it below?

library(dplyr)
library(prophet)    

df <- data_frame(ds = seq(as.Date("2017/01/01"), as.Date("2019/01/01"), "month"), 
                     a = rnorm(n = 25, mean = 100000, sd = 7500), 
                     b = rnorm(n = 25, mean = 100000, sd = 7500), 
                     c = rnorm(n = 25, mean = 100000, sd = 7500))

The columns a, b and c are sales numbers for each product. And, I would like to run forecasting for all 3 products and the total sales in next 12 periods.

So, I tidy up the dataframe, then do group forecasting.

d1 <- df %>%
        gather(key = "prod", value = "y", a:c) %>%
        nest(-prod) %>%
        mutate(m = map(data, prophet)) %>%
        mutate(future = map(m, make_future_dataframe, period = 12, freq = "month")) %>%
        mutate(forecast = map2(m, future, predict)) %>%
        mutate(p = map2(m, forecast, plot))

The output looks like this:

# A tibble: 3 x 6
  dept  data              m             future                forecast               p       
  <chr> <list>            <list>        <list>                <list>                 <list>  
1 a     <tibble [25 x 2]> <S3: prophet> <data.frame [37 x 1]> <data.frame [37 x 16]> <S3: gg>
2 b     <tibble [25 x 2]> <S3: prophet> <data.frame [37 x 1]> <data.frame [37 x 16]> <S3: gg>
3 c     <tibble [25 x 2]> <S3: prophet> <data.frame [37 x 1]> <data.frame [37 x 16]> <S3: gg>

Then, I create the plots manually, and rearrange them using grid.arrange

gridExtra::grid.arrange(d1$p[[1]], d1$p[[2]], d1$p[[3]]

Are there any ways to do it quicker and automatically?

Upvotes: 0

Views: 900

Answers (2)

Victor HDC
Victor HDC

Reputation: 555

Single plot: Using cowplot

cowplot::plot_grid(plotlist =d1$p)

Multiple plots: Using purrr::walk function:

walk(.x=d1$p,.f=gridExtra::grid.arrange)

.x is the first parameter and .f is the function. If you need more than one parameter then you should use walk2 (2) or pwalk (3+)

Upvotes: 0

RLave
RLave

Reputation: 8364

I couldn't find a way to pipe the call to grid.arrange, but you can use do.call to avoid manually selecting plots.

do.call(gridExtra::grid.arrange, d1$p)

Example:

library(dplyr)
library(prophet)  
library(tidyr)
library(purrr)

df <- data_frame(ds = seq(as.Date("2017/01/01"), as.Date("2019/01/01"), "month"), 
                 a = rnorm(n = 25, mean = 100000, sd = 7500), 
                 b = rnorm(n = 25, mean = 100000, sd = 7500), 
                 c = rnorm(n = 25, mean = 100000, sd = 7500))



d1 <- df %>%
  gather(key = "prod", value = "y", a:c) %>%
  nest(-prod) %>%
  mutate(m = map(data, prophet)) %>%
  mutate(future = map(m, make_future_dataframe, period = 12, freq = "month")) %>%
  mutate(forecast = map2(m, future, predict)) %>%
  mutate(p = map2(m, forecast, plot))

Upvotes: 3

Related Questions