The project consists of an interactive data visualization that enables the detection of patterns in the cultural representation of Tokyo on Instagram. A database consisting of 40,851 images was obtained using a system of crawling data from the web. The system collects images shared on Instagram that have labeled Tokyo by hashtags in 47 different languages. An Image Caption Generation model based on (Vinyals, Toshev, Bengio, & Erhan, 2016) was applied to each of these images. From the resulting descriptions, the system obtains main “motifs” (words with the highest semantic content) by using Natural Language Processing (Bird & Loper, 2004). When a viewer interacts with the system it is displayed the images with the recurring motifs in each of the languages, the image descriptions, and the hashtags associated with them.

​

References

• Bird, S., & Loper, E. (2004). NLTK: The Natural Language Toolkit. ACL Home Association for Computational Linguistics, 214–217. Retrieved SEP 09, 2020, from https://www.aclweb.org/anthology/P04-3031

• Vinyals, O., Toshev, A., Bengio, S., & Erhan, D. (2016). Show and Tell: Lessons learned from the 2015 MSCOCO Image Captioning Challenge. IEEE Transaction on Pattern Analysis and Machine Intelligence, Vol. XX (No. XX ). doi:DOI: http://dx.doi.org/10.1109/TPAMI.2016.2587640

Title: The unveiled city
Year: 2018-2019
Technique: Interactive data visualization. PC, TV (58”, 4K), webcam, mechanical arm, custom table (wood and pipes), printed graphic (framed).

Dimensions: Variable, minimum: 3m x 2m.
Edition: 2 + 2PA. There is an unreleased high scale version of the project.
Shows:

• June 7, 2019. As part of the solo show <H1> Hello Tokyo </ h1>. ARTnSHELTER Gallery, Tokyo, Japan
   

Direction and concept: Yonlay Cabrera, Luis Diago
Implementation: Yonlay Cabrera, Luis Diago, Gustavo Viera, Antonio Serrano
Support: Tama Art University Graduate School, Information Design Department; Meiji Institute for Advanced Study of Mathematical Science (MIMS).