Understanding basic Plots behavior


I’m working my way through basic Julia exercises in a course. I have no problem with most of it but I haven’t yet grasped the odd way plots seem to work. Here’s a basic fn for calculating the steady state values of a logistic map:

function steady_state_values(r)
    b = 0.25
    for i in 1:400
        b = r * b * (1-b)
    attractor_values = zeros(150)
    for i in 1:150
        b = r * b * (1-b)
        attractor_values[i] = b
    return attractor_values

I want to vary the parameter r (along the x axis) and plot the corresponding steady state values along y. I’ve tried variations of this:

using Plots
first = true
for r = 2.9 : 0.001 : 4
    y = steady_state_values(r)
    x = fill(r, length(y))
    if first
        plot(x=x, y=y)
        first = false
        plot!(x=x, y=y)

Multiple calls to plot don’t seem to work–they result in no output-- so I plot the first array then call plot! subsequently. Even this doesn’t work. Can someone explain the logic? I come from matplotlib where you call figure() to begin, then any number of plotting operations, then show().
Plots.jl doesn’t seem to work that way. I’ve gone through a Plots.jl notebook but it didn’t clarify.



plot(...) creates and returns a plot object. plot!(plt,...) modifies the plot object plt and returns it. If plt is not specified, the last plot touched is used.

In Juno and Jupyter, a plot is shown when it’s the output of the cell/line that was just run. It’s just like with any return value, except instead of showing a textual representation, it’s graphical. Your code ended in a for loop, which has the value nothing, so nothing is shown or printed.
To see the plot, you can do

using Plots
plt = plot()
for r = ...
    plot!(x, y)  # or plot!(plt,x,y)

The last line is to display the plot plt.


Are you doing this in Jupyter or a .jl file? It can matter when for loops are involved… my suggestion is to stay in Jupyter for this sort of stuff unless your loops in a function.


I was doing this directly in Jupyter.


Thanks! That’s it exactly.

(In a Python notebook the plotting is a side effect, and the value doesn’t matter. Hence my confusion.)