Install and setup Plots with PyPlot backend using system Python in for Julia. Then builds system-specific Julia image for faster performance in the future.

julia-install-pyplot-windows.jl

# Note: Both Julia and Python must be of the same architecture type. If Julia
#       is 64-bit, so must your Python distribution.
#
#       I tested this with a clean Julia install, by installing Julia and
#       deleting "C:\Users\<username>\.julia".
#
#       Julia:  v6.1
#       Python: v3.6.3 for Windows
#

test_after_install = true

Pkg.update()

# Must point to actual python.exe, not just python dir
if is_linux()
    ENV["PYTHON"] = "python3"
elseif is_windows()
    //ENV["PYTHON"] = "C:/Users/Alex/AppData/Local/Programs/Python/Python36/python.exe"  
    
    ENV["PYTHON"] = "EDITME"                                # <<---- EDIT THIS LINE 
end
    
python = ENV["PYTHON"]

run(`$python -m pip install --upgrade --user pip`)
run(`$python -m pip install --upgrade --user numpy`)
run(`$python -m pip install --upgrade --user matplotlib`)
run(`$python -m pip install --upgrade --user scipy`)


Pkg.add("PyCall")
Pkg.build("PyCall")

Pkg.add("PyPlot")
Pkg.add("Plots")
Pkg.build("Plots")


# Build system image
# From: #7 http://www.stochasticlifestyle.com/7-julia-gotchas-handle/

if is_windows()
    Pkg.add("WinRPM"); 
    using WinRPM
    WinRPM.install("gcc", yes=true)
    WinRPM.install("winpthreads-devel", yes=true)
end

include(joinpath(dirname(JULIA_HOME),"share","julia","build_sysimg.jl")); build_sysimg(force=true)

if test_after_install
    using PyPlot
    display(plot(rand(100)))
end