1
$\begingroup$

Suppose we have a 2D distribution of shops in an area. For example, here is a heatmap density plot of (latitude, longitude) location of shops in a province: enter image description here

Now suppose that we "zoom into" the distribution and consider only a tiny window $w:([x_1,x_2],[y_1,y_2])$ inside the whole range of x,y; so that we can have a better insight into how data points (shops) are arranged in the small local neighbourhoods.

The data distribution INSIDE each random window is obviously a subset of the entire data. For example, if we plot shops with latitude between $[42, 42.3]$ and longitude between $[-76, -75.5]$, we get this distribution: (visualized with scatterplot)

enter image description here

In the statistics community, what is the actual name of these "zoomed in" parts of a data distribution?

I want to report that the data distribution is skew in general, but quite uniform in "those small subsets". What are those distributions inside the windows called?

May I say, for example, "micro-distribution" or "partial distribution"?

Improve this question
$\endgroup$
6
  • 1
    $\begingroup$ Notice that you can always edit a question, not need to delete and ask new question. In fact, we generally discourage deleting and re-asking, in vast majority of cases it will yield a moderator reaction. $\endgroup$
    – Tim ♦
    Jun 3 '20 at 10:17
  • $\begingroup$ @Tim yep, thanks for the hint & sorry for the repost; but I realized that my initial post was so badly written that it drifted all comment to irrelevant subjects (such as the type of the charts, the scale of the features, etc). $\endgroup$
    – Ali
    Jun 3 '20 at 12:01
  • 1
    $\begingroup$ Detail view of ... $\endgroup$
    – wolfies
    Jun 3 '20 at 15:10
  • $\begingroup$ If your goal is to talk about some property, you can just say that the property applies "locally" or "in this region" $\endgroup$
    – user20160
    Jun 3 '20 at 18:15
  • $\begingroup$ @user20160, I am indexing a multidimensional data which splits the data into a grid. I am interested in the data distribution inside each grid cell. Is "local distribution" a common term in statistics? $\endgroup$
    – Ali
    Jun 4 '20 at 16:21
1
$\begingroup$

This would be the conditional distribution, conditional on $x\in[x_1,x_2]$ and $y\in[y_1,y_2]$.

Your phrasing of “zoom in” is descriptive and might be worth including, something like, “We also zoom in on the window to show the conditional distribution...”

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Agree with second paragraph, but I'd be reluctant to use "conditional" for describing data manipulations. Conditional probability is a probabilistic term, here we are talking about filtering the data. I wouldn't mix those. $\endgroup$
    – Tim ♦
    Jun 3 '20 at 11:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.