Emulating Google Sheets smoothed line chart with ggplot2

Question

I'm trying to emulate a Google Sheets smoothed line chart with ggplot2. It draws a smoothed line through all data points.

plot in Google Sheets: The lines above and below is my hardcoded confidence intervals.

data:

library(tidyverse)
data <- tibble(
    date = seq.Date(as.Date("2018-12-01"), as.Date("2018-12-20"), by = "days"),
    var = c(0.329,0.348,0.349,0.355,0.382,0.363,0.340,0.359,0.336,0.358,0.398,0.389,0.389,0.390,0.383,0.343,0.352,0.415,0.397,0.430),
    lower = c(0.311,0.330,0.330,0.336,0.364,0.345,0.321,0.342,0.319,0.341,0.384,0.375,0.375,0.374,0.369,0.329,0.337,0.400,0.382,0.417),
    upper = c(0.347,0.366,0.368,0.374,0.400,0.381,0.358,0.377,0.354,0.375,0.413,0.404,0.403,0.405,0.397,0.357,0.368,0.430,0.412,0.444))

I almost get the same curve with loess. But it gives me a warning and a strange sharp-looked areas at some points.

code:

ggplot(data, aes(x = seq_along(date), y = var)) + 
    geom_point(size = 2, color = "blue", alpha = 0.2) +
    # geom_smooth(method = "lm", formula = y ~ splines::bs(x, 20), se = F) +
    # stat_smooth(method = "gam", formula = y ~ s(x, k = 19), se = F) +
    stat_smooth(method = "loess", span = 0.2, se = F) +
    theme_classic() + theme(axis.line = element_line(size = 0.5, colour = "grey80"))

warning:

Warning messages:
1: In simpleLoess(y, x, w, span, degree = degree, parametric = parametric,  :
    Chernobyl! trL>n 20
2: In simpleLoess(y, x, w, span, degree = degree, parametric = parametric,  :
    Chernobyl! trL>n 20
3: In sqrt(sum.squares/one.delta) : NaNs produced

R plot:

Not quite a duplicate, but this should help: stackoverflow.com/q/35205795/5325862 Google Sheets might be doing LOESS behind the scenes, but instead some other way of calculating smooth curves between points — camille
– camille, Commented Dec 22, 2018 at 15:40

pogibas · Accepted Answer · 2018-12-22 15:45:44Z

Just for the visualisation you can try awesome ggalt package by @hrbrmstr.

ggalt::geom_xspline draws x-spline.

library(ggalt)
library(tidyverse)

data <- tibble(
    date = seq.Date(as.Date("2018-12-01"), as.Date("2018-12-20"), by = "days"),
    var = c(0.329,0.348,0.349,0.355,0.382,0.363,0.340,0.359,0.336,0.358,0.398,0.389,0.389,0.390,0.383,0.343,0.352,0.415,0.397,0.430),
    lower = c(0.311,0.330,0.330,0.336,0.364,0.345,0.321,0.342,0.319,0.341,0.384,0.375,0.375,0.374,0.369,0.329,0.337,0.400,0.382,0.417),
    upper = c(0.347,0.366,0.368,0.374,0.400,0.381,0.358,0.377,0.354,0.375,0.413,0.404,0.403,0.405,0.397,0.357,0.368,0.430,0.412,0.444))

ggplot(data, aes(seq_along(date), var)) + 
    geom_point(size = 2, color = "blue", alpha = 0.2) +
    geom_xspline() +
    theme_classic() + 
    theme(axis.line = element_line(size = 0.5, colour = "grey80"))

Collectives™ on Stack Overflow

Emulating Google Sheets smoothed line chart with ggplot2

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related