Inconsistent Scatter Plots

User 2986 | 1/9/2016, 3:35:49 PM

I am taking a course on Coursera that is using GraphLab Create. Strangely I got three different scatter plots from the same set of data for the same code. The code was-

<code> sales.show(view = 'Scatter Plot', x = 'sqft_living', y = 'price') </code>

The instructor got this plot-

I got this plot when the canvas target was the browser-

And I got this one when the canvas target was iPython notebook-

I should also mention in none of the cases above scatter plots for the max and min 'sqft_living' were made.

Why is this happening?

I think I should also mention that the size of the data set was 21613 rows x 21 columns. So that would make 21613 pairs of (sqft_living, price)

Comments

User 2986 | 1/9/2016, 5:04:53 PM

Are the points to be plotted randomly chosen? That is the only explanation I can think of.


User 2535 | 1/12/2016, 7:01:09 PM

Hi @Shabbir,

Yes, it's sampling internally, using a different random seed each time it plots. We are looking to address this in a future release.

Thanks. Jon


User 2986 | 1/13/2016, 12:11:13 PM

Thank you for the information @jon.