dartmouth-pbs / psyc161 Goto Github PK

I'm running into trouble with the creation of a scatterplot with kiva data in section 8.15. act_kiva_graph_2 (Create a well scaled scatter plot with X and Y axes using num_lenders_total and loan_amount).
The turtle is painstakingly slow, regardless of the speed parameter, and drawing the loan_amount-num_lenders_total scatterplot fails due to a RunTime Error after the turtle finished the second dot.

This error occurs consistently: it happens in the sandbox, it happens if I build runestone locally from the pbs-psych161-wi19 branch, it happens if I build runestone locally from the current state of runestone/ac101.

The same code runs flawless locally in an ipython instance using Python 3.5:

import turtle
sc = turtle.Screen()
adina = turtle.Turtle()
adina.speed(10)
sc.setworldcoordinates(0, 0, max(num_lenders_total), max(loan_amount))

adina.penup()

for i in range(len(num_lenders_total)):
    adina.goto(num_lenders_total[i], loan_amount[i])
    adina.stamp()

A workaround for me was to scale the loan_amount variable, as this one was clearly responsible for giving my poor turtle a hard time (very long ways to walk for it). The following code works AND is also able to adjust the speed:

loan_amount_scale = [i/500 for i in loan_amount]

import turtle
sc = turtle.Screen()
adina = turtle.Turtle()
adina.speed(0)
sc.setworldcoordinates(0, 0, max(num_lenders_total), max(loan_amount_scale))

adina.penup()

for i in range(len(num_lenders_total)):
    adina.goto(num_lenders_total[i], loan_amount_scale[i])
    adina.stamp()

But I doubt that this is an optimal solution...

The idea is inspired by #2 . Although original problem is technical, and we could/should solve it technically, its slowness also opens the door to demonstrate advantages of actually thoughtful design and algorithms. In this case sorting could be of initial help and the ultimate order of the screen exploration would be the NP complete https://en.wikipedia.org/wiki/Travelling_salesman_problem .