Relationship between pixel and record on plotting a graph

The code you've posted doesn't run. At least there is no x to take a len() of. So probably it should look like this:

import matplotlib.pyplot as plt

y = [1]*19200
x = range(0,19200)

plt.scatter(x,y)
plt.show()

But that way it just draws a line.

Even if you'd have a monitor with 19200 pixels you'll still see a line with a dot in every pixels of it.

So I suppose you've actually meant something like this:

import matplotlib.pyplot as plt
import numpy as np

L = 19200

y = [1]*L
x = [v*(i+1) for i, v in enumerate(np.random.randint(0,2,L))]

plt.scatter(x,y,alpha=0.5)
plt.show()

On a really big monitor it would look like this:

And of course you can't show separate points in one pixel, but you can show how many points fit into one particular pixel. Just add another dimension to your plot.

As your plot is just a line and you don't really utilize y axis you can use it as an extra dimension:

x = np.random.randint(0,2,size=(L,10))
y = np.sum(x,axis=1)
x = range(0,L)

plt.fill_between(x,y, alpha=0.5)
plt.show()

It will give:

Height of a bar represents number of points in a pixel.

Or if you really want a line you can use color as an extra dimension:

x = np.random.randint(0,2,size=(L,10))
colors = [[(1,1,1),(0,0,1),(0,0.5,1),(0,1,1),(0,1,0.5),(0,1,0),(0.5,1,0),(1,1,0),(1,0.5,0),(1,0,0)][v] for v in np.sum(x,axis=1)]
# or if your prefer monochrome
# colors = [(1,0,0,v/10) for v in np.sum(x,axis=1)]
x = range(0,L)
y = [1]*L

plt.scatter(x,y, c=colors, s=3)
plt.show()

It gives:

Monochrome version:

You can play with it here

^{As images are even smaller that 1920 I played with the L to make them more vivid on a random data.}

Update:

There are 3 points in the first group of pixels. And there is one point in the second.

Image:

How many points do you see in the third group?

If you have a good eyesight and a big screen you'll see that there are 2 point.

But what if you have a poor eyesight, a small screen, or if you'll just step away from the screen. Can you still tell how many points are there? Yes, you can!

^{To some extent, of course. If you'll stand in 1Km from the screen you'll probably won't be able to see the screen itself :)}

But how can you tell? By the weight of the group - it looks lighter than the first one and darker than the second.

Now, show the next image to someone and tell that the first group has tree pixels. Then ask: how many pixels are in other groups?

Image:

They'll probably tell you that there are 2 and 1 pixels in those group. But that's not true. There are same number of pixels. The only difference is that those pixels have different color.

So, it doesn't really matters how many pixels you'll draw. What matters is how they perceived.

But more than that... You say "pixel", but is it a dot? No! In most cases there are 3 dots of different color.

So if you see a red pixel you can be sure that there is one dot lighten up. If you see yellow - there are 2 dots lighten up. Etc. Judging by the color you can even say exactly which of the dots constituting the pixel are highlighted.

But again - does it really matter? If you'll just say: "this particular color means (0,1)(1,1)(2,1), and this particular color means (3,1)(4,1), etc." people will understand your plot regardless of monitors and their resolutions.

But, again, more than that: when you draw a pixel on your monitor it is not even a single physical pixel and not just 3 dots. You monitor has a maximum resolution of 8192*8192... so there is more that 8 physical pixels for a one logical on resolution 1920*1080. And this actually gives more than 16 physical pixels for one logical. So can you

Put 10 x records into just one box?

...as you can see, the "box" is quite big. You can put 16 records into it. Physically. And logically you can add even more.