How to make forecast charts in R?

Question

I have a model that makes predictions and would like to plot the observed series and then the predictions, according to this graph:

I tried to use the geom_smooth along with the geom_line, but unsuccessfully, keeps returning error.

Code I am using:

#remotes::install_github("joachim-gassen/tidycovid19")
library(tidyverse)
library(openxlsx)
library(tidycovid19)

previsoes <- read.xlsx("https://raw.githubusercontent.com/gabrielrvsc/COVID19_outputs/master/forecasts_20200513.xlsx")
previsoes$Data <- convertToDate(previsoes$Data)

updates <- download_merged_data(cached = TRUE)

updates %>%
  filter(iso3c %in% "BRA", date > "2020-03-20") %>%
  ggplot() +
  geom_line(aes(x = date, y = confirmed)) +
  geom_line(previsoes, aes(x = Data, y = Previsao_Casos))

Answer 1

I tested your code here and the error that appeared was

Error: Mapping should be created with aes() oraes_()`.

To fix this, you need to declare 'date'

geom_line(data = previsoes, aes(x = Data, y = Previsao_Casos))

NOTE. Your chart will not look like the one you ask because the data in 'forecasts' only has May, from 14/05. There is a fault in the line because it lacks day 14/05 in the 'updates'. It jumps because of the NA

Removed 1 Row(s) containing Missing values (geom_path).

You’ll have to treat that NA for it to join the lines

OBS.: I redid the code by creating data frames to better organize, fix plotting flaws and get to plot subtitles, etc.

library(tidyverse)
library(openxlsx)
library(tidycovid19)

previsoes <- read.xlsx("https://raw.githubusercontent.com/gabrielrvsc/COVID19_outputs/master/forecasts_20200513.xlsx")
previsoes$Data <- convertToDate(previsoes$Data)

updates <- download_merged_data(cached = TRUE)

df1 <- updates %>%
   filter(iso3c %in% "BRA", date > "2020-03-20") %>%
   select(date, confirmed) %>% 
   rename(Data = date, Casos = confirmed) %>% 
   mutate(Tipo = "Confirmados")

df2 <- previsoes %>% 
   select(Data, Previsao_Casos) %>% 
   rename(Casos = Previsao_Casos) %>% 
   mutate(Tipo = "Previsto")

df3 <- rbind(df1, df2)

options(scipen = 999)

grafico <- df3 %>%  
   ggplot() +
   geom_line(aes(x = Data, y = Casos, linetype = Tipo)) +
   geom_vline(xintercept = as.Date("2020-05-14"), lty = 2) +
   scale_x_date(date_breaks = "7 days", date_labels = "%d/%b") +
   geom_ribbon(data = previsoes, aes(x = Data, ymin = Previsao_Casos-Previsao_Casos_Variacao, ymax = Previsao_Casos+Previsao_Casos_Variacao), alpha = 0.3)