geom_smooth lm和geom_line没有绘制相同的线

Question

I am drawing a linear line with geom_smooth method lm and geom_line based on a fitted lm model, however, they don't draw the same curve.

Which one is right ? Is there any error in my code ?

Below the code :

sales <- structure(list(NET_PRICE_TTC = c(59.988, 54.448, 54.99, 59.9895, 
59.990666, 59.994, 54.454, 55.707428, 59.992, 49.3044, 59.988, 
60.504, 44.9064, 40.336, 55.5855, 57.989599, 56.983999, 59.989714, 
59.989714, 60.501, 60.504, 56.18057, 57.265091, 60.504, 59.9925, 
47.077, 59.988, 60.504, 60.504, 53.97, 60.504, 59.988, 57.847714, 
56.467999, 56.211199, 59.989999, 60.504, 57.48, 52.683, 49.898117, 
60.504, 49.896, 59.997, 48.9, 57.49, 60.504, 49.483999, 59.9928, 
60.504, 60.504, 60.504, 58.488, 50.55857, 47.136, 54.4512, 60.504, 
59.988, 60.4968, 49.896, 48.9, 58.483999, 59.988, 60.5064, 52.459999, 
49.901142, 59.991, 53.988, 51.900799, 60.375, 60.42857, 59.988, 
52.504999, 59.988, 55.488, 59.988, 49.8975, 59.9892, 59.988, 
60.5064, 56.718, 59.988, 59.989714, 59.623636, 59.988, 47.547157, 
60.50057, 60.3492, 59.988, 59.991, 59.9928, 59.988, 59.988, 60.504, 
59.988, 48.01457, 43.4904, 48.9, 59.988, 48.901333, 44.994, 56.990768, 
56.2395, 59.988, 60.498, 59.989091, 59.988, 59.989999, 48.881777, 
55.491, 59.9904, 59.986285, 56.443636, 52.9395, 59.988, 59.989714, 
59.988, 45.86516, 51.310956, 60.504, 56.182285, 60.499636, 48.900666, 
56.574666, 52.1415, 51.159, 59.988, 59.827999, 57.262908, 59.988, 
60.5016, 59.988, 60.445333, 58.74, 57.48, 59.991, 50.349332, 
60.282545, 52.941, 60.4992, 44.374736, 59.748, 59.988, 55.705714
), QTY = c(10, 6, 12, 16, 18, 4, 12, 14, 6, 20, 16, 12, 10, 6, 
16, 30, 6, 14, 14, 8, 22, 14, 22, 12, 16, 48, 12, 2, 8, 20, 2, 
6, 28, 12, 30, 12, 10, 8, 8, 34, 6, 10, 8, 18, 24, 10, 24, 10, 
12, 6, 6, 6, 28, 56, 10, 4, 16, 10, 16, 28, 6, 10, 10, 24, 14, 
16, 10, 30, 8, 14, 2, 24, 12, 4, 4, 16, 20, 18, 10, 16, 4, 14, 
22, 10, 38, 14, 20, 24, 8, 10, 16, 2, 10, 16, 56, 40, 24, 12, 
18, 4, 26, 16, 4, 4, 22, 10, 24, 54, 16, 10, 14, 22, 16, 10, 
14, 14, 62, 46, 8, 14, 22, 36, 18, 16, 16, 2, 18, 22, 10, 10, 
8, 18, 32, 6, 8, 18, 22, 8, 10, 38, 8, 8, 14)), .Names = c("NET_PRICE_TTC", 
"QTY"), class = "data.frame", row.names = c(NA, -143L))

lin <- lm(QTY ~ NET_PRICE_TTC, sales)

lin_intercept = coef(lin)["(Intercept)"]
lin_price_coef = coef(lin)["NET_PRICE_TTC"]

demand_curve = data.frame(price_point = seq(45, 60, by = 1))
demand_curve %<>% mutate(quantity = lin_intercept + lin_price_coef * price_point)


ggplot(sales, aes(x = QTY, y = NET_PRICE_TTC)) +
  geom_point() +
  geom_line(data = demand_curve, aes(x = quantity, y = price_point)) +
  geom_smooth(
    method = lm,
    color = "black",
    se = FALSE,
    fullrange = TRUE
  )

Answer 1

There is an error in your code. In your regression you take QTY as a function of NET_PRICE_TTC , while in your ggplot call you set x=QTY and y=NET_PRICE_TTC (and the same goes for the geom_line call).

Flipping x and y in the ggplot call solves it. Here I coloured the lm fitted line blue.

ggplot(sales, aes(y = QTY, x = NET_PRICE_TTC)) +
  geom_point() +
  geom_line(data = demand_curve, aes(y = quantity, x = price_point), col='blue', size=2) +
  geom_smooth(
    method = lm,
    color = "black",
    se = FALSE,
    fullrange = TRUE
  )