Parallel approaches have been proposed in wide areas for improving the system performance. The proposal introduces three approaches to process statistical data tabulation. First, task parallelism to exploit the use of multi-cores in shared memory architecture is developed according to a bushy tree query plan for decomposing tasks. Each decomposed task is allocated to a processor core such that the workloads among processor cores are nearly balanced. Secondly, data parallelism to utilize available interconnected PCs is developed according to master-slave paradigm. Data distribution adopts a simple Round Robin approach to achieve workload balance and no data dependency among PCs. Lastly, hybrid parallelism is proposed to combine the former approaches for optimizing the use of available processor cores in interconnected PCs. The experimental results shows considerably good performance in terms of parallel time execution, speed up and efficiency.