Abstract

Investment, either by domestic or foreign investors, is one of the important components of the economic growth as it contributes directly onto the national product. As a long-term commitment, it is important for the policy makers to conduct a very well-defined strategy by looking around any evidence that can potentially project the future of the investment condition. With this study, we propose an alternative approach in measuring the investment direction by combining high frequency structured data (real time gross settlement/RTGS transaction data) and unstructured data (news data). Unlike previous approaches that leverage either investor’s survey discussion forum or traditional national statistics office, our approach offers more timely and flexible information. Generally, the indicator we construct consists of two kind of indicators by considering the magnitude of temporal: leading indicator and prompt indicator. The leading indicator is constructed by applying text mining into textual data from online news data to proxy the willingness of the investors. On the other hand, the prompt indicator is constructed by using big data analytics onto the RTGS data to proxy the realisation of the investment. On the result, we found that the leading indicator is likely to forecast the realisation of the investment for the next some quarters, while prompt indicator is likely to nowcast the current direction of the investment. These indicators have potential to be used as initial proxy for portraying investment condition in Indonesia.