Recently data stream mining is new emerging field of data. It is different from traditional static data. In stream data, data is changing dynamically. They have characteristics like timing preference, data distribution changes constantly with time, data flows in and out with speed, amount of data is enormous and immediate response is required. So, to preserve the privacy during data stream mining many privacy preserving techniques have been proposed. Many existing techniques for privacy preserving are suitable for traditional static database but not suitable for dynamic data. So in data stream privacy preservation is an important issue. To achieve the privacy and balancing the accuracy we will use geometric data transformation (perturbation) technique. This paper present survey about geometric data perturbation technique for privacy preserving data mining.