It's a case of definitions. Resistance is defined by R = V/I. So you just divide V by I.
For a conductor which obeys Ohm's law, a graph of V against I is a straight line through the origin, so ΔV/ΔI at any point gives you exactly the same thing as V/I.
For non-ohmic devices (filament lamps, diodes and so on) the graph of V against I is not straight, so V/I is not a constant. [Nor does V/I usually equal ΔV/ΔI at a point.] Since V/I is not a constant for a non-ohmic device, the concept of resistance (defined as V/I) is far less useful for such a device: one might as well go back to the I against V curve itself, when doing calculations.
As the last poster pointed out, there are certain devices for which one is concerned with changes in V associated with small changes in I, for changes centred on a particular point on the V – I curve. In that case what we're interested in isn't V/I but really is ΔV/ΔI. This quantity is sometimes called 'slope resistance'. The reciprocal is 'slope conductance'.
For displacement – time (x – t) graphs, velocity is defined as [the limit as Δt approaches zero of] Δx/Δt because that's what we're interested in. [How fast was the car going when it crashed? There's very little interest in knowing x/t, the mean velocity since the journey started.
In general, things are defined as they are defined, because they're interesting and/or important when defined that way.