Transactional data of library loans play an important role for knowing how library materials are actually used, and for many years, the data have been statistically analyzed for obtaining suggestions that lead to improvement of library services and operations. Since data mining and so-called big data have drawn attention, it seems that analysis of library use data is moving on to a next stage. This paper reviews some topics of it from the past to presence. Specifically, after overviewing briefly bibliometrics, data mining and big data, loan data analysis for weeding, collection assessment and recommendation of books are described with some examples. Gathering and applying of library use data other than loan data are also discussed.

The University of Electro-Communications Library built an innovative space named Ambient Intelligence Agora (AIA), in which seminars, workshops, and other meetings can be held, in cooperation with AI research. In this space, various data reflecting the state of the learning environment are acquired by installing various sensors, and they can be analyzed and visualized by AI research using deep learning machine. Here, we will present the current situation and future image of AIA from three different viewpoints, i.e. ambient environment, active learning and AI research, and robotic intellectual interaction, and discuss how AIA contributes to realization of the next-generation library.

Among college/university libraries, it has been very common to record, compile and analyze the entry history of library patrons to improve library services, though the analysis was not helpful enough to clarify the facts such as seat availability, length of staying time, or behaviors of the library patrons. In this paper, we introduce our experiments using IoT devices to collect and analyze data for behavioral research mainly focused on searching library items. BLE (Bluetooth Low Energy) beacons have been installed in the library building, while the research participants carry smartphones customized for the experiments. Data obtained through the smartphones are sent to the beacons, and integrated to reveal the patrons’ searching behavior and path-finding in the building.

As an example of forecasting technology trends in advanced technologies, we show how we discovered new product areas using printing technologies not currently commercialized, but likely to be applied in future. Focusing on the difference in properties between patents including technical information close to commercialization and journals including fundamental scientific research results that can lead to inventions and product developments in the future, we extracted “a field where the number of journals recently issued trend upward though patent applications are sluggish” as “a field likely to develop into commercialization in the future”. After further text mining and manpowered screening analyses of the patents and journals information in the extracted fields to refine, we identified the intended product fields.