Lucid was called a dataflow language at one point, but is more accurately an 'eduction' language. The key difference is that Lucid relies on demand-driven evaluation, wheras dataflow semantics entirely relies on the more push-based semantics of data becoming available to fire an operation. I suppose you could call it 'tagged-token demand-driven dataflow', but its semantics are much different from other dataflow languages such as SISAL, VAL or Id.
https://news.ycombinator.com/item?id=2714745