Can the ship perform more calculations per second since it is much larger and is not confined to the size of an average human? Or is the ship actually a distributed network of many computers specialized for specific tasks, thus making the two difficult to objectively compare?
Interestingly, there is an episode where the ship and Data directly interact with one another on a digital scale - A Fistful of Datas.
In this episode, Georgi and Data attempt to connect Data to the Enterprise's main computer, so that in the event of an emergency Data could take full control of the Enterprise.
Their experiment goes remarkably well, until Data tries to access the ship's secondary systems. Data detects an overload in his positronic matrix and shuts the experiment down, but the damage is done in the form of a Texas-Taklin' Data, as well as Data being represented in several parts of the ship (most notably, as several characters on the Holodeck).
Up until that point though, Data was able to access individual parts of the system and operate them. The sudden surge, however, suggests that some systems may be too complex for him to handle. And the fact that he was accessing systems one at a time also lends credit to suggest the Enterprise's computer is larger and more complex than Data's brain can handle all at once.
This may also indicate a limit on Data's ability to multi-task as well as the ship computer, or that there is simply an incompatability between the two. Unfortunately, there is no way to tell, because the series does not do a particularly deep analysis into the ship's data processing power.
In TNG S03E06, "Booby Trap," Geordi suggests that not even Commander Data could perform as many adjustments per second as the ship's computer (in the form of a holographic projection of one of the original Enterprise's warp core designers).
I am sure that I remember reading in the ST tech manual that all federation ships central computer processor core are held inside of a static warp bubble (hence making them FTL processors), secondary interface sub systems run at less than FTL, (pads control panels etc)
It was mentioned on Voyager by its own computer that it has transluminal processing. This implies from the wording alone that the main computer processor is FTL based...
And since most (or practically all) technology on a Federation ship is subspace based, even computers would effectively be using subspace principles.
It never made any sense that the Federation was unable to create its own Androids. Heck, given the rate of exponential progression, androids with more processing power than Data will be possible shortly in reality... Star Trek missed the mark on 'exponential development' aspect especially when it had various aliens for example being better at translation than it... that's just impossible. No organic entity would be able to process things faster than a computer... a computer simply needs a proper set of algorithms to work with... and while those used in Trek seemed advanced, they also seemed outdated in comparison to the hardware's capabilities.