Due to the impact of Covid-19, distance education has been approved by the Ministry of Education and implemented all over the country as an alternative during the pandemic. In order to provide for this new situation in which students’ self-study is deemed to increase exponentially, we have tried to create a databank, Japanese e-Learning Website and Data Collection (JeLD). The collection consists of over 70 databases, repositories and academic and educational sources selected from students’ most visited websites. JeLD aims to increase the efficiency of online data collection from educational media sources, Japanese language systems and academic websites used by universities in Japan. This paper aims to show the results of the application of JeLD in NCKU and of the AI analysis of essays collected from corpus and e-Learning website in JeLD, e.g. the corpus of Taiwanese learner of Japanese (CTLJ) and Japanese e-Learning based on CTLJ. Furthermore, the latter helps improve teaching methods by comparing different characteristics of writing between native Japanese and non-natives.