Where are all those values I need?
You can ask NumPy.
Just ask `np.where()`?
Let's see how this function works through an example…
👇🪡🧵
#python #numpy
1/
You can ask NumPy.
Just ask `np.where()`?
Let's see how this function works through an example…
👇🪡🧵
#python #numpy
1/
A teacher asks his students to take a test twice, a few weeks apart.
2/
2/
He thinks this is a fair system to use:
• If the difference in scores between the first and second tests is less than five marks, he'll keep the result of the first test
• Otherwise, he'll take the highest mark
3/
• If the difference in scores between the first and second tests is less than five marks, he'll keep the result of the first test
• Otherwise, he'll take the highest mark
3/
If you're not convinced that's a fair test, neither am I.
But that's what the teacher settled on…
Let's help the teacher write code to work this out using NumPy:
4/
But that's what the teacher settled on…
Let's help the teacher write code to work this out using NumPy:
4/
The arrays `first_test` and `second_test` contain the marks for 5 students.
Let's look at how you use `np.where()` bit by bit.
The function takes three arguments.
In the code snippet above, the three arguments are shown on separate lines.
5/
Let's look at how you use `np.where()` bit by bit.
The function takes three arguments.
In the code snippet above, the three arguments are shown on separate lines.
5/
The first argument is the condition you want to use to decide which value you want to keep.
In this case, you're checking whether the difference between the first and second test scores is less than 5.
6/
In this case, you're checking whether the difference between the first and second test scores is less than 5.
6/
If this condition is `True`, the value from the second argument is used.
In this example, if the difference is less than 5 marks, the score from the first test is used.
7/
In this example, if the difference is less than 5 marks, the score from the first test is used.
7/
However, if the condition in the first argument in `False`, the value used is the one in the third argument.
In this example, this is the highest of the two scores.
8/
In this example, this is the highest of the two scores.
8/
`np.maximum()` returns an array where each element is the maximum value from the corresponding elements in the arrays passed as arguments.
It's not the same as `np.max()` which returns the maximum value from one array.
9/
It's not the same as `np.max()` which returns the maximum value from one array.
9/
The second and third arguments you use in `np.where()` can be either arrays of the same shape and size, or ones which can be broadcast using NumPy broadcasting rules.
Here's an example:
10/
Here's an example:
10/
`a` is an array of shape (7, ). For broadcasting purposes, this is the same as (1, 7)
`b` has shape (7, 1)
The output from `np.where()` is a (7, 7) array
11/
`b` has shape (7, 1)
The output from `np.where()` is a (7, 7) array
11/
In the first row of the output, the values in `a` are compared to the first value of `b` which is 3.
When the value of `a` is less than 3, then `a`'s value is used. Otherwise 3 is used.
The same applies for the other rows.
12/
When the value of `a` is less than 3, then `a`'s value is used. Otherwise 3 is used.
The same applies for the other rows.
12/
So, if you're wondering where to find the data you need, you may want to ask NumPy…
`np.where()` is one of the many tools you have in NumPy that allow you to explore your data.
13/
`np.where()` is one of the many tools you have in NumPy that allow you to explore your data.
13/
You'll get lots of NumPy stuff if you follow on @s_gruppetta_ct…
…and Matplotlib, and general Python fundamentals, and…
… examples that take the `turtle` module to the limit!
14/
…and Matplotlib, and general Python fundamentals, and…
… examples that take the `turtle` module to the limit!
14/
https://twitter.com/s_gruppetta_ct/status/1533898133989511168
• • •
Missing some Tweet in this thread? You can try to
force a refresh
