Exploratory data analysis

Abstract

Exploratory data analysis techniques are used to detect trends and other statistical characteristics in nine stream flow time series at both the annual and monthly levels. For convenience of interpretation, the output from the analysis is displayed graphically, along with some numerical results from appropriate statistical tests. As well as providing indications and statistical tests of trends, non-normal behaviour and auto correlated behaviour in flow sequences, exploratory data analysis may be used to place any particular response or collection of responses in context against the range of observed values, thus indicating periods of unusual flow conditions. © 1991 Taylor & Francis Group, LLC.