Towards Better Policy and Practice Using Real-Time Data

Abstract

Can Search Engines (SE) queries and Tweets help predict economic activity? Can Social Networking Services (SNS) help understand time use and help predict online activity? Clearly, those two questions address different issues, but both emphasize the use of (near) real-time data to refine and improve the comprehension of our Society. Indeed, in a fast changing world such as the one we live in, it appears more and more crucial for policy and practice to be aware of what is going on in the very present in view of proposing responsive and relevant solutions in the near future. One very innovative and promising way to achieve this is by using online real-time data. Real-time data denotes information that is delivered immediately after collection, meaning there is no delay or extremely small delay in the time-line of information provided. For example, thanks to the Twitter API , it is nowadays possible to easily gather data on what people are doing/thinking right now. Besides, tools like Google Trends provide a very simple way to collect hourly, daily or weekly data on the volume of queries related to various keywords. Now, those tools enable researchers to get a fantastic amount of data on an extremely large set of topics. For instance, an economist could use SE queries related to a specific industry (e.g. cars) in a specific part of the world (e.g. France) to refine his seasonal auto-regressive model for this industry in this country. Similarly, it could be interesting to analyze SE queries and tweets making reference to ‘job interviews’ or to online job platforms such as ‘monster.com’ to help predict global and local unemployment rates. Finally, data gathering on the individuals’ online activity, that is, for example, the number of tweets and re-tweets per units of time (e.g. minutes) or the number of Facebook status updates per units of time could help understand when people are the most active online and, performed together with a contents’ semantic analysis, could help understand for what purpose. This short paper aims at proposing methods to enhance current policy and practice. The first section is interested in enhancing current forecasting methods for macroeconomic activity by taking into account real-time data. The second section proposes and discusses methods using real time data to examine the time use of individuals or groups of individuals. Finally, conclusion takes place in the last section.