One context I use decimal code points for is Windows Alt codes. For example, the degree sign (°) is \u00B0, but typed on Windows as Alt+0176 (digits typed on the numpad).
I consider this a failing of Windows, given the widespread use of hex code points everywhere else, but it would be (very slightly) more convenient not to need to memorize two separate ways of typing special characters that are common for my work.
I’m -1 on the proposal, as I might use it if it existed, but don’t think it justifies the breakage with existing behavior on \d. If a separate PEP could motivate a larger change to error on all invalid Unicode escapes, then I’d likely become +1 after the transition period.