Forecasting my weight with R

[This article was first published on Econometrics and Free Software, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

I’ve been measuring my weight almost daily for almost 2 years now; I actually started earlier, but not as consistently. The goal of this blog post is to get re-acquaiented with time series; I haven’t had the opportunity to work with time series for a long time now and I have seen that quite a few packages that deal with time series have been released on CRAN. In this blog post, I will explore my weight measurements using some functions from the {tsibble} and {tibbletime} packages, and then do some predictions with the {forecast} package.

First, let’s load the needed packages, read in the data and convert it to a tsibble:

library("tidyverse")
library("readr")
library("forecast")
library("tsibble")
library("tibbletime")
library("mice")
weight <- read_csv("https://gist.githubusercontent.com/b-rodrigues/ea60679135f8dbed448ccf66a216811f/raw/18b469f3b0720f76ce5ee2715d0f9574b615f170/gistfile1.txt") %>% 
    as_tsibble()
## Parsed with column specification:
## cols(
##   Date = col_date(format = ""),
##   Poids = col_double()
## )
## The `index` is `Date`.

You can read more about {tsibble} here. Here, I use {tsibble} mostly for the next step, which is using the function fill_na() on the tsibble. fill_na() turns implicit missing values into explicit missing values. These are implicit missing values:

          Date Poids
1   2013-01-01 84.10
2   2013-01-04 85.60

and this is the same view, but with explicit missing values:

          Date Poids
1   2013-01-01 84.10
2   2013-01-02 NA
3   2013-01-03 NA
4   2013-01-04 85.60

This is useful to do, because I want to impute the missing values using the {mice} package. Let’s do this:

weight <- weight %>% 
    fill_na()

imp_weight <- mice(data = weight) %>% 
    mice::complete("long")
## 
##  iter imp variable
##   1   1  Poids
##   1   2  Poids
##   1   3  Poids
##   1   4  Poids
##   1   5  Poids
##   2   1  Poids
##   2   2  Poids
##   2   3  Poids
##   2   4  Poids
##   2   5  Poids
##   3   1  Poids
##   3   2  Poids
##   3   3  Poids
##   3   4  Poids
##   3   5  Poids
##   4   1  Poids
##   4   2  Poids
##   4   3  Poids
##   4   4  Poids
##   4   5  Poids
##   5   1  Poids
##   5   2  Poids
##   5   3  Poids
##   5   4  Poids
##   5   5  Poids

Let’s take a look at imp_weight:

head(imp_weight)
##   .imp .id       Date Poids
## 1    1   1 2013-10-28  84.1
## 2    1   2 2013-10-29  84.4
## 3    1   3 2013-10-30  83.5
## 4    1   4 2013-10-31  84.1
## 5    1   5 2013-11-01  85.6
## 6    1   6 2013-11-02  85.2

Let’s select the relevant data. I filter from the 11th of July 2016, which is where I started weighing myself almost every day, to the 31st of May 2018. I want to predict my weight for the month of June (you might think of the month of June 2018 as the test data, and the rest as training data):

imp_weight_train <- imp_weight %>% 
    filter(Date >= "2016-07-11", Date <= "2018-05-31")

In the next lines, I create a column called imputation which is simply the same as the column .imp but of character class, remove unneeded columns and rename some other columns (“Poids” is French for weight):

imp_weight_train <- imp_weight_train %>% 
    mutate(imputation = as.character(.imp)) %>% 
    select(-.id, -.imp) %>% 
    rename(date = Date) %>% 
    rename(weight = Poids)

Let’s take a look at the data:

ggplot(imp_weight_train, aes(date, weight, colour = imputation)) +
    geom_line() + 
    theme(legend.position = "bottom")

This plots gives some info, but it might be better to smooth the lines. This is possible by computing a rolling mean. For this I will use the rollify() function of the {tibbletime} package:

mean_roll_5 <- rollify(mean, window = 5)
mean_roll_10 <- rollify(mean, window = 10)

rollify() can be seen as an adverb, pretty much like purrr::safely(); rollify() is a higher order function that literally rollifies a function, in this case mean() which means that rollifying the mean creates a function that returns the rolling mean. The window argument lets you decide how smooth you want the curve to be: the higher the smoother. However, you will lose some observations. Let’s use this functions to add the rolling means to the data frame:

imp_weight_train <- imp_weight_train %>% 
    group_by(imputation) %>% 
    mutate(roll_5 = mean_roll_5(weight),
           roll_10 = mean_roll_10(weight))

Now, let’s plot these new curves:

ggplot(imp_weight_train, aes(date, roll_5, colour = imputation)) +
    geom_line() + 
    theme(legend.position = "bottom")
## Warning: Removed 20 rows containing missing values (geom_path).

ggplot(imp_weight_train, aes(date, roll_10, colour = imputation)) +
    geom_line() + 
    theme(legend.position = "bottom")
## Warning: Removed 45 rows containing missing values (geom_path).

That’s easier to read, isn’t it?

Now, I will use the auto.arima() function to train a model on the data to forecast my weight for the month of June. However, my data, imp_weight_train is a list of datasets. auto.arima() does not take a data frame as an argument, much less so a list of datasets. I’ll create a wrapper around auto.arima() that works on a dataset, and then map it to the list of datasets:

auto.arima.df <- function(data, y, ...){

    y <- enquo(y)

    yts <- data %>% 
        pull(!!y) %>% 
        as.ts()

    auto.arima(yts, ...)
}

auto.arima.df() takes a data frame as argument, and then y, which is the column that contains the univariate time series. This column then gets pulled out of the data frame, converted to a time series object with as.ts(), and then passed down to auto.arima(). I can now use this function on my list of data sets. The first step is to nest the data:

nested_data <- imp_weight_train %>% 
    group_by(imputation) %>% 
    nest() 

Let’s take a look at nested_data:

nested_data
## # A tibble: 5 x 2
##   imputation data              
##   <chr>      <list>            
## 1 1          <tibble [690 × 4]>
## 2 2          <tibble [690 × 4]>
## 3 3          <tibble [690 × 4]>
## 4 4          <tibble [690 × 4]>
## 5 5          <tibble [690 × 4]>

nested_data is a tibble with a column called data, which is a so-called list-column. Each element of data is itself a tibble. This is a useful structure, because now I can map auto.arima.df() to the data frame:

models <- nested_data %>% 
    mutate(model = map(data, auto.arima.df, y = weight))

This trick can be a bit difficult to follow the first time you see it. The idea is the following: nested_data is a tibble. Thus, I can add a column to it using mutate(). So far so good. Now that I am “inside” the mutate call, I can use purrr::map(). Why? purrr::map() takes a list and then a function as arguments. Remember that data is a list column; you can see it above, the type of the column data is list. So data is a list, and thus can be used inside purrr::map(). Great. Now, what is inside data? tibbles, where inside each of them is a column called weight. This is the column that contains my univariate time series I want to model. Let’s take a look at models:

models
## # A tibble: 5 x 3
##   imputation data               model      
##   <chr>      <list>             <list>     
## 1 1          <tibble [690 × 4]> <S3: ARIMA>
## 2 2          <tibble [690 × 4]> <S3: ARIMA>
## 3 3          <tibble [690 × 4]> <S3: ARIMA>
## 4 4          <tibble [690 × 4]> <S3: ARIMA>
## 5 5          <tibble [690 × 4]> <S3: ARIMA>

models is a tibble with a column called model, where each element is a model of type ARIMA.

Adding forecasts is based on the same trick as above, and we use the forecast() function:

forecasts <- models %>% 
    mutate(predictions = map(model, forecast, h = 24)) %>% 
    mutate(predictions = map(predictions, as_tibble)) %>% 
    pull(predictions) 

I forecast 24 days (I am writing this on the 24th of June), and convert the predictions to tibbles, and then pull only the predictions tibble:

forecasts
## [[1]]
## # A tibble: 24 x 5
##    `Point Forecast` `Lo 80` `Hi 80` `Lo 95` `Hi 95`
##  *            <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1             71.5    70.7    72.3    70.2    72.8
##  2             71.5    70.7    72.4    70.3    72.8
##  3             71.5    70.6    72.3    70.1    72.8
##  4             71.5    70.6    72.4    70.1    72.9
##  5             71.4    70.5    72.4    70.0    72.9
##  6             71.5    70.5    72.4    70.0    72.9
##  7             71.4    70.5    72.4    69.9    72.9
##  8             71.4    70.4    72.4    69.9    72.9
##  9             71.4    70.4    72.4    69.9    72.9
## 10             71.4    70.4    72.4    69.8    73.0
## # ... with 14 more rows
## 
## [[2]]
## # A tibble: 24 x 5
##    `Point Forecast` `Lo 80` `Hi 80` `Lo 95` `Hi 95`
##  *            <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1             71.6    70.8    72.3    70.3    72.8
##  2             71.6    70.8    72.5    70.3    72.9
##  3             71.5    70.6    72.4    70.2    72.9
##  4             71.5    70.6    72.5    70.1    72.9
##  5             71.5    70.5    72.5    70.0    73.0
##  6             71.5    70.5    72.5    70.0    73.0
##  7             71.5    70.5    72.5    69.9    73.0
##  8             71.5    70.4    72.5    69.9    73.1
##  9             71.5    70.4    72.5    69.8    73.1
## 10             71.4    70.3    72.6    69.7    73.1
## # ... with 14 more rows
## 
## [[3]]
## # A tibble: 24 x 5
##    `Point Forecast` `Lo 80` `Hi 80` `Lo 95` `Hi 95`
##  *            <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1             71.6    70.8    72.4    70.4    72.8
##  2             71.5    70.7    72.4    70.2    72.8
##  3             71.5    70.6    72.4    70.2    72.9
##  4             71.5    70.6    72.4    70.1    72.9
##  5             71.5    70.5    72.4    70.0    72.9
##  6             71.5    70.5    72.4    70.0    73.0
##  7             71.5    70.5    72.5    69.9    73.0
##  8             71.4    70.4    72.5    69.9    73.0
##  9             71.4    70.4    72.5    69.8    73.0
## 10             71.4    70.4    72.5    69.8    73.1
## # ... with 14 more rows
## 
## [[4]]
## # A tibble: 24 x 5
##    `Point Forecast` `Lo 80` `Hi 80` `Lo 95` `Hi 95`
##  *            <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1             71.5    70.8    72.3    70.3    72.8
##  2             71.5    70.7    72.4    70.3    72.8
##  3             71.5    70.7    72.4    70.2    72.8
##  4             71.5    70.6    72.4    70.1    72.9
##  5             71.5    70.6    72.4    70.1    72.9
##  6             71.5    70.5    72.5    70.0    73.0
##  7             71.5    70.5    72.5    69.9    73.0
##  8             71.5    70.4    72.5    69.9    73.0
##  9             71.4    70.4    72.5    69.8    73.1
## 10             71.4    70.3    72.5    69.8    73.1
## # ... with 14 more rows
## 
## [[5]]
## # A tibble: 24 x 5
##    `Point Forecast` `Lo 80` `Hi 80` `Lo 95` `Hi 95`
##  *            <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1             71.5    70.8    72.3    70.3    72.8
##  2             71.5    70.7    72.4    70.3    72.8
##  3             71.5    70.7    72.4    70.2    72.8
##  4             71.5    70.6    72.4    70.1    72.9
##  5             71.5    70.6    72.4    70.1    72.9
##  6             71.5    70.5    72.4    70.0    73.0
##  7             71.5    70.5    72.5    69.9    73.0
##  8             71.5    70.4    72.5    69.9    73.0
##  9             71.4    70.4    72.5    69.8    73.1
## 10             71.4    70.3    72.5    69.8    73.1
## # ... with 14 more rows

So forecasts is a list of tibble, each containing a forecast. Remember that I have 5 tibbles, because I imputed the data 5 times. I will merge this list of data sets together into one, but before I need to add a column that indices the forecasts:

forecasts <- map2(.x = forecasts, .y = as.character(seq(1, 5)), 
     ~mutate(.x, id = .y)) %>% 
    bind_rows() %>% 
    select(-c(`Lo 80`, `Hi 80`))

colnames(forecasts) <- c("point_forecast", "low_95", "hi_95", "id")

Let’s take a look again at forecasts:

forecasts
## # A tibble: 120 x 4
##    point_forecast low_95 hi_95 id   
##             <dbl>  <dbl> <dbl> <chr>
##  1           71.5   70.2  72.8 1    
##  2           71.5   70.3  72.8 1    
##  3           71.5   70.1  72.8 1    
##  4           71.5   70.1  72.9 1    
##  5           71.4   70.0  72.9 1    
##  6           71.5   70.0  72.9 1    
##  7           71.4   69.9  72.9 1    
##  8           71.4   69.9  72.9 1    
##  9           71.4   69.9  72.9 1    
## 10           71.4   69.8  73.0 1    
## # ... with 110 more rows

I now select the true values for the month of June. I also imputed this data, but here I will simply keep the average of the imputations:

weight_june <- imp_weight %>% 
    filter(Date >= "2018-06-01") %>% 
    select(-.id) %>% 
    group_by(Date) %>% 
    summarise(true_weight = mean(Poids)) %>% 
    rename(date = Date)

Let’s take a look at weight_june:

weight_june
## # A tibble: 24 x 2
##    date       true_weight
##    <date>           <dbl>
##  1 2018-06-01        71.8
##  2 2018-06-02        70.8
##  3 2018-06-03        71.2
##  4 2018-06-04        71.4
##  5 2018-06-05        70.9
##  6 2018-06-06        70.8
##  7 2018-06-07        70.5
##  8 2018-06-08        70.1
##  9 2018-06-09        70.3
## 10 2018-06-10        71.0
## # ... with 14 more rows

Let’s repeat weight_june 5 times, and add the index 1 to 5. Why? Because I want to merge the true data with the forecasts, and having the data in this form makes things easier:

weight_june <- modify(list_along(1:5), ~`<-`(., weight_june)) %>% 
    map2(.y = as.character(seq(1, 5)), 
         ~mutate(.x, id = .y)) %>% 
    bind_rows()

The first line:

modify(list_along(1:5), ~`<-`(., weight_june)) 

looks quite complicated, but you will see that it is not, once we break it apart. modify() modifies a list. The list to modify is list_along(1:5), which create a list of NULLs:

list_along(1:5)
## [[1]]
## NULL
## 
## [[2]]
## NULL
## 
## [[3]]
## NULL
## 
## [[4]]
## NULL
## 
## [[5]]
## NULL

The second argument of modify() is either a function or a formula. I created the following formula:

~`<-`(., weight_june)

We all know the function <-(), but are not used to see it that way. But consider the following:

a <- 3
`<-`(a, 3)

These two formulations are equivalent. So these lines fill the empty element of the list of NULLs with the data frame weight_june. Then I add the id column and then bind the rows together: bind_rows().

Let’s bind the columns of weight_june and forecasts and take a look at it:

forecasts <- bind_cols(weight_june, forecasts) %>% 
    select(-id1)

forecasts
## # A tibble: 120 x 6
##    date       true_weight id    point_forecast low_95 hi_95
##    <date>           <dbl> <chr>          <dbl>  <dbl> <dbl>
##  1 2018-06-01        71.8 1               71.5   70.2  72.8
##  2 2018-06-02        70.8 1               71.5   70.3  72.8
##  3 2018-06-03        71.2 1               71.5   70.1  72.8
##  4 2018-06-04        71.4 1               71.5   70.1  72.9
##  5 2018-06-05        70.9 1               71.4   70.0  72.9
##  6 2018-06-06        70.8 1               71.5   70.0  72.9
##  7 2018-06-07        70.5 1               71.4   69.9  72.9
##  8 2018-06-08        70.1 1               71.4   69.9  72.9
##  9 2018-06-09        70.3 1               71.4   69.9  72.9
## 10 2018-06-10        71.0 1               71.4   69.8  73.0
## # ... with 110 more rows

Now, for the last plot:

ggplot(forecasts, aes(x = date, colour = id)) +
    geom_line(aes(y = true_weight), size = 2) + 
    geom_line(aes(y = hi_95)) + 
    geom_line(aes(y = low_95)) + 
    theme(legend.position = "bottom")

The true data fall within all the confidence intervals, but I am a bit surprised by the intervals, especially the upper confidence intervals; they all are way above 72kg, however my true weight has been fluctuating around 71kg for quite some months now. I think I have to refresh my memory on time series, because I am certainly missing something!

If you found this blog post useful, you might want to follow me on twitter for blog post updates.

To leave a comment for the author, please follow the link and comment on their blog: Econometrics and Free Software.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)