Big Data Based Cash Flow Prediction in Liquidity Planning

Currently, cash flow forecasting within an enterprise requires a consolidation of all subsidiary companies to obtain the planned cash flow figures for future periods. The consolidation-process is very time consuming as it requires persons of each subsidiary company to submit the cash flow figures. The motivation of this master thesis is the simplification of this process by providing estimated cash flow figures. The goal of this thesis is to predict cash flows of an entire group. The approach to achieve this goal is the prototypical implementation of a predictive model to predict cash flows for each company. Based on a literature review an appropriate predictive model which is scalable was chosen. The resulting prototype has been implemented in R according to the Cross-Industry Standard Process for Data Mining (CRISP-DM). Furthermore, the predictive model was successfully integrated into an SQL Server environment which enables the prediction of cash flows using simple SQL queries. The resulting cash flows are stored in the database where it is possible to integrate the result into a liquidity planning software to assist users. The resulting prototype has been developed in a general approach so that it can be used as a basis for cash flow prediction of various branches.