Historical Big Data
Historical big data is a project about seamless analysis of the environment and the society from the past and the present based on various records written by humans.
Old books in Edo period was written by Kuzushiji (old Japanese characters), and modern Japanese people can rarely read the characters without effort. Then, instead of humans, can AI read Kuzushiji after training? We release Character Shape dataset to the world as a training data for machine learning to tackle the grand challenge of "Kuzushiji vs. AI" with the collective power of people.
On November 9, 1867, the restoration of imperial rule symbolized the end of Edo Period. 150 years have passed since then, and now is the time to revive the information space of Edo, using open data about the 260 years of Edo period, and taking advantage of the state-of-the-art technologysuch as artificial intelligence (AI).
Bukan Complete Collection
The project aims at analyzing comprehensively the collection of "Bukan" books, which is the best seller through the 200 years of Edo period, and constructing core information platform about Edo period in terms of human and geospatial information about Daimyo (lords) and Shogunate government.
North China Railway Archive
A research database on North China Railway Company by linking company's promotional stock photographs with its transportation network, and studying the activities of the company from the theme and location of photographs.
Digital Silk Road
Digital humanities research project about creating digital archives of cultural heritage based on collaboration between informatics and humanities.
Memory Platform / Memorygraph
Memorygraph is a new photographic technique to create the layer of memories, and the project aims to develop the Memorygraph app to use it for field work of cultural heritage, tourism, and recovery from disasters.
A project that aims at integrating geographic information science (GIS) and natural language processing (NLP) to develop a geo-tagging system that transforms text to maps automatically.