Remove Outdated NY Taxi benchmark

[alamb]

Which issue does this PR close?

Part of #7052

Rationale for this change

1. This is not a standard benchmark
2. The benchmark code currently checked into the repo is a single low cardinality grouping query, a scenario that is already well covered by both tpch and clickbench
3. I also think the benchmark has atrophied (because, for example, the https://www.nyc.gov/site/tlc/about/tlc-trip-record-data.page now only supplies parquet where the benchmark still has references to CSV data)

FWIW the query is:

SELECT passenger_count, MIN(fare_amount), MAX(fare_amount), SUM(fare_amount) FROM tripdata GROUP BY passenger_count");

What changes are included in this PR?

Remove the NY taxi benchmark code and entry in readme

Are these changes tested?

Are there any user-facing changes?

[alamb]

@andygrove / @Dandandan do you know of anyone using the nytaxi benchmark?

[Dandandan]

Fine to remove by me

[jackwener]

Make sense to me