Time series (ts) function in R doesn't count correctly

Question

When using the ts function in R to create a time series per month, the the totals per time unit do not seem to add up correctly. For example, the value for Jan-2014 should be well over 100, but the output says it's 3 (see below). Does anyone know what's wrong with my code?

Used code:

set.seed(1)
df <- data.frame(date_rec = seq(as.Date("2014-01-01"),as.Date("2015-12-31"), by=1),
                 number = sample(c(1:10),size = 730, replace=TRUE))

df_ts <-  ts(df$number, frequency=12, start=c(2014,1), end=c(2015,12))
df_ts

Result for df_ts

    Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2014   3   4   6  10   3   9  10   7   7   1   3   2
2015   7   4   8   5   8  10   4   8  10   3   7   2

Answer 1

Another alternative using aggregate before ts:

library(tidyverse)
# Group by month and aggregate the numbers
monthly_data <- aggregate(number ~ year(date_rec) + month(date_rec), data = df, FUN = sum)

# Create a time series object
ts(monthly_data$number, frequency = 12, start = c(2014, 1))

   year(date_rec) month(date_rec) number
1            2014               1    188
2            2015               1    143
3            2014               2    159
4            2015               2    148
5            2014               3    189
6            2015               3    198
7            2014               4    157
8            2015               4    158
9            2014               5    169
10           2015               5    195
11           2014               6    171
12           2015               6    176
13           2014               7    195
14           2015               7    172
15           2014               8    181
16           2015               8    175
17           2014               9    187
18           2015               9    196
19           2014              10    167
20           2015              10    187
21           2014              11    162
22           2015              11    204
23           2014              12    189
24           2015              12    165

Answer 2

Summarising the daily data by month before creating the ts object:

library(tidyverse)

set.seed(1)
df <- data.frame(
  date_rec = seq(as.Date("2014-01-01"), as.Date("2015-12-31"), by = 1),
  number = sample(c(1:10), size = 730, replace = TRUE)
)

df |> 
  mutate(date_rec = round_date(date_rec, unit = "month")) |> 
  summarise(number = sum(number), .by = date_rec) |> 
  ts(frequency = 12, start = c(2014, 1), end = c(2015, 12))
#>          date_rec number
#> Jan 2014    16071     88
#> Feb 2014    16102    196
#> Mar 2014    16130    166
#> Apr 2014    16161    179
#> May 2014    16191    148
#> Jun 2014    16222    160
#> Jul 2014    16252    180
#> Aug 2014    16283    185
#> Sep 2014    16314    155
#> Oct 2014    16344    160
#> Nov 2014    16375    160
#> Dec 2014    16405    164
#> Jan 2015    16436    200
#> Feb 2015    16467    147
#> Mar 2015    16495    157
#> Apr 2015    16526    165
#> May 2015    16556    189
#> Jun 2015    16587    138
#> Jul 2015    16617    150
#> Aug 2015    16648    157
#> Sep 2015    16679    180
#> Oct 2015    16709    192
#> Nov 2015    16740    179
#> Dec 2015    16770    181

^{Created on 2024-03-18 with reprex v2.1.0}