line plot on bar chart with ggplot2

Question

I would like to present the data visualisation which includes line chart and bar chart. Could you guys tell me what kind of mistakes I have made? I guess if it because the moving weighted average caused this issue?

rollmean_covid  <- covid_ts1  %>% filter(Nation == "England") %>% select(date, daily_pos_num) %>% 
    mutate(pos_num01 = rollmean(daily_pos_num, k = 3, fill = NA),
           pos_num02 = rollmean(daily_pos_num, k = 5, fill = NA),
           pos_num03 = rollmean(daily_pos_num, k = 7, fill = NA),
           pos_num04 = rollmean(daily_pos_num, k = 10, fill = NA),
           pos_num05 = rollmean(daily_pos_num, k = 14, fill = NA))

rollmean_covid_metric <- rollmean_covid %>% gather(metric, number, pos_num01:pos_num05)
rollmean_covid_metric %>% filter(metric == "pos_num01") %>% 
  ggplot() + 
  geom_line(aes(date, number), col = "Blue") +
  geom_col(aes(date, number), fill = "orange", alpha = 0.7)

edited1: the dataset overview is provided bellow.

# A tibble: 10 x 7
   date       daily_pos_num pos_num01 pos_num02 pos_num03 pos_num04 pos_num05
   <date>             <dbl>     <dbl>     <dbl>     <dbl>     <dbl>     <dbl>
 1 2020-01-30             2    NA          NA      NA          NA      NA    
 2 2020-01-30             0     0.667      NA      NA          NA      NA    
 3 2020-01-31             0     0           0.4    NA          NA      NA    
 4 2020-01-31             0     0           0       0.286      NA      NA    
 5 2020-02-01             0     0           0       0           0.3    NA    
 6 2020-02-01             0     0           0       0.143       0.1    NA    
 7 2020-02-02             0     0           0.2     0.143       0.1     0.286
 8 2020-02-02             0     0.333       0.2     0.143       0.2     0.143
 9 2020-02-03             1     0.333       0.2     0.143       0.2     0.143
10 2020-02-03             0     0.333       0.2     0.286       0.2     0.143

@AllanCameron I expect the line is on the top of bar, but I fail to plot the line over the bar — James Huang
– James Huang, Commented Jul 22, 2020 at 17:36

dave-edison · Accepted Answer · 2020-07-22 19:22:32Z

2

You have multiple observations for one date, and by default the position = "stack" for geom_col so that's why the bars are taller. You probably want position = "identity", this should work:

rollmean_covid_metric %>% 
    filter(metric == "pos_num01") %>% 
    ggplot(aes(date, number)) + 
    geom_line(color = "blue") +
    geom_col(position = "identity", fill = "orange", alpha = 0.7)

Edit to add: as @chemdork123 points out in the comments, if one date has multiple values this probably won't give desired results. In general, the best solution for these types of problems is to munge your data into the correct shape before piping it into a ggplot call.

edited Jul 22, 2020 at 19:22

answered Jul 22, 2020 at 17:44

dave-edison

3,7469 silver badges20 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

chemdork123 Over a year ago

This answer +1. OP, your filter() command prior to plotting is giving you duplicate data. Check the data preparation prior to the ggplot call. Setting position="identity" will work to fix your plot, but you will still be plotting two copies of geom_col on top of one another. If they are duplicates... no huge deal, you'll just have darker bars. If they are not duplicates, your results will still look wrong. Without the original dataset, it's difficult to see where you're going wrong going from rollmean_covid to rollmean_covid_metric, but the answer lies within there.

James Huang Over a year ago

Thanks for your help@DiceboyT @chemdork123 , this not only justify my problem but also let me more understand the logics of ggplot2

Collectives™ on Stack Overflow

line plot on bar chart with ggplot2

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related