Imagine you have a function that takes in one variable. In math, we usually write is as f(x). If we want to find its slope at any given point, we take its derivative. That function is f'(x) (read as f prime of x) but what if the function is based on two variables or more? Now if we want to find its slope as one variable changes but the rest are constant, we don't use a derivative. We use a partial derivative.
For a function, say f(x, y), its partial derivatives are ∂f/∂x and ∂f/∂y, each with respect to its own variable (x and y respectively)
To calculate a partial derivative, let's do an example: f(x, y) = 2x^(2) - xy + 5.
∂f/∂x with this example will be d/dx (2x^(2) - xy + 5) (where we treat y as a constant)
differentiating this the normal way gives 4x - y. That's the partial derivative.
Similarly, ∂f/∂x = d/dx (2x^(2) - xy + 5) (where we treat y as a constant) so, the partial derivative with respect to y is -x.
Why use partial derivatives? Whenever you want to measure change in temperature, calculate profit, predict weather patterns, partial derivatives make it a lot easier. They show up in many real world problems, and I believe developing a solid grasp of them is essential.