The invention is related to a method and a monitoring system for sample-analysis of data comprising a multitude of data packets. Defined parent population numbers (N) of data packets are sampled by an out of N sampling and classified or vice versa. Classification is done by classification rules to create sample-flow-groups of specific flows. Each sample-flow-group consists of a sample-flow-quantity of data packets having a sample-flow-size. According to the invention in each sample-flow-group sample-flow-variances and sample-flow ratios defined by the sample-flow-quantity in proportion to the sample number (n) are determined. Furthermore, an estimated flow size defined by an estimated sum of data sizes of data packages in a flow of the parent population number of data packets and its variance are calculated. For the calculation the parent population number, the sample number and at least one of the sample-flow-ratios, the sample-flow-mean-sizes and the sample-flow-variances are used, comprising the step of determination of at least one of parent population flow-ratios parent population flow-mean-sizes and parent population flow-variances for each flow in the parent population number of data packets.