CODEWITHSUNDEEP
rggplot2
DaveRGP
DaveRGP

Reputation: 1480

Is it possible to bypass stat_smooth when using geom_smooth?

I have a data set that I have built the following plot around: enter image description here

The plot is fed from a data set included at the bottom of this post, and is made from the following ggplot2 code:

    ggFacetProfile <- ggplot(sub, aes(group = iMoYr))  + 
    geom_line(aes(x= iHrMi, y = trimAv)) + 
    facet_grid(off ~ iMoYr, scales = "free") +
    ggtitle("Typical Half Hourly Profiles") + 
    xlab("Time") + ylab("Energy (kWh)")

Here I am plotting the values trimAv (effectively average values) over iHrMi (effectively hour and minute). This is on a facet of off by iMoYr (effectively is the process on of off, and different months over the year).

The data table already has the effective smoothed value ranges calculated in it, under the heading minEcl and maxEcl. I would like to be able to use geom_smooth to represent this data on the graph as boundaries of the shape made by the function geom_smooth, however I haven't been able to locate a way to bypass calling stat_smooth.

My closest attempt so far is to include:

+ geom_smooth(aes(x= iHrMi, y = trimAv, ymin = minEcl, ymax = maxEcl))

However, this is coerced into a loess smoothing, apparently due to the size of the data, which looks like this:

enter image description here

Is it possible to feed geom_smooth specific pre-calculated values, or am I trying to use geom_smooth in very much the wrong way? It seems incongruous that the other geom_ arguments in ggplot2 are so adaptable and this seems so rigid.

The head and tail of the data source (a data table) is included below for structure purposes:

          iDate            off       trimAv   trimStD minEcl maxEcl   iMoYr iHrMi
   1: 2013-08 00:00     Production 136.52273 37.300389   76.4  218.4 2013-08 00:00
   2: 2013-08 00:30     Production 136.14091 36.117819   80.3  217.7 2013-08 00:30
   3: 2013-08 01:00     Production 133.92500 32.808662   76.9  213.3 2013-08 01:00
   4: 2013-08 01:30     Production 139.20476 37.929480   77.1  221.5 2013-08 01:30
   5: 2013-08 02:00     Production 137.82857 36.422042   74.9  221.0 2013-08 02:00
  ---                                                                             
1148: 2014-07 22:30 Non-Production  50.51250  3.025812   47.1   56.3 2014-07 22:30
1149: 2014-07 23:00 Non-Production  49.88571  2.066743   47.0   52.6 2014-07 23:00
1150: 2014-07 23:30 Non-Production  49.94286  2.318661   46.5   52.5 2014-07 23:30
1151: 2014-07 00:00 Non-Production  50.85714  2.860569   47.9   54.9 2014-07 00:00
1152: 2014-07 00:30 Non-Production  50.72857  4.181194   47.6   59.1 2014-07 00:30

If I can include the source data in a better/ more suitable form please let me know in the comments.

Upvotes: 2

Views: 269

Answers (1)

Julian
Julian

Reputation: 791

Maybe you are looking for geom_ribbon. 

ggFacetProfile <- ggplot(sub, aes(group = iMoYr))  + 
  geom_line(aes(x= iHrMi, y = trimAv)) + 
  facet_grid(off ~ iMoYr, scales = "free") +
  ggtitle("Typical Half Hourly Profiles") + 
  xlab("Time") + ylab("Energy (kWh)") +
  geom_ribbon(aes( ymin = minEcl, ymax = maxEcl))

Upvotes: 1

Related Questions