Oddly, I was reading a paper by Gregory Schopen tonight that addresses this particular issue- the preponderance of suttas set in Savatti.
As far back as Caroline Rhys Davids, scholars have suggested that the prominence of Savatti might have been a product of redaction- she suggested the possibility that Savatti was the first major collection point or library of sutras. Schopen’s research into rules for the redaction of texts found in the Mulasarvastivada Vinaya suggests that, over time, discourses whose location was unknown or forgotten may have been ‘assigned’ a location in Savatti due to the presence of two major lay supporters, the king Pasenadi and the layman Anathapindika.
Of course, their presence could also account for why the Buddha spent so much time there in the first place- the presence of the Jetavana monastery, where the Buddha supposedly spent as many as 20 rainy season retreats, would naturally lead to the production of suttas set in that location.
While the specific text he sights is specific to the Mulasarvastivada Vinaya and doesn’t have a known parallel in the Theravada tradition, the locations used in the Pali sutras, the Madhyamika Agama (Chinese parallel of the Majjhima Nikaya), and the Pali jataka collection are all described as potentially reflecting the application of these redaction rules.
For instance, one rule states that when an ancient location is unknown, the city of Varanasi (Benares) should be chosen and the king Brahmadatta should be used for an unknown king. Varanasi is the setting for the majority of the Pali jataka in the PTS edition, and events are described as taking place in the reign of Brahmadatta. In some cases, this combination of location and ruler are specified at the start of the story even when they have no bearing on the rest of the narrative. In one text, the name of another king is used for an ancient ruler of Benares, but later in the text he is addressed as ‘Brahmadatta’ without explanation.
Naturally, the discovery of such a rule isn’t conclusive (since these locations would have to have been prominent in the first place for such rules to have been composed) and Schopen estimates that the rule is from relatively late in the Mulasarvastivada Vinaya. However, it does seem to stem from an era when some sutras were still being written down locally for the first time, due to the presence for rules about recording a sutra as text. It also suggests that sutras being converted from purely oral to written form may have been a long-term process, rather than an event.