Toshiba Digital Solutions collaborates with DATAFLUCT to Deliver a Machine
Learning Solution that Optimizes Store Visitors Prediction

April 21, 2021
Toshiba Digital Solutions Corporation
DATAFLUCT, Inc.

KAWASAKI and TOKYO―Toshiba Digital Solutions Corporation (TDSL) and DATAFLUCT, Inc. (DATAFLUCT) today announced a machine learning solution that optimizes the prediction of the number of visitors to a store. The solution is possible through the combination of TDSL’s GridDB Cloud^*1, a cloud data platform capable of delivering real-time analysis and DATAFLUCT's DATAFLUCT cloud terminal., which enables utilization of advanced machine learning without the need for an expert. This solution makes it possible to quickly and easily predict the number of visitors to a store, which is difficult to realize without technical expertise and a complex data infrastructure. The two companies began collaborating in May 2020 through Toshiba OPEN INNOVATION PROGRAM 2020^*2,3 organized by Toshiba Corporation with the aim of creating new businesses. This collaboration results in a solution that will be brought to the market and sold by DATAFLUCT. 

Until now, most of the store visitors’ predictions are based on human intuition and experience. However, as we enter an era where intuition and experience become less relevant, it is difficult to respond to major changes in social conditions, consumer behavior, as well as the different characteristics of a store such as its surrounding environment and customer base. One of the solutions to this problem is data utilization. By using data to predict the number of visitors and customer segments for a store, accurate decisions can be made in regards to the amount of merchandise ordered, the timing of displays, etc., to maximize sales. However, if the accumulated data is inadequately analyzed, the structure of the data is disjointed and unusable for analysis, or if there is a shortage of data scientists, it becomes impossible to create a highly accurate prediction model. In order to solve these problems, GridDB Cloud and DATAFLUCT cloud terminal. are brought together to create a highly accurate prediction model using in-house and external data. This makes it possible to quickly and easily predict the number of visitors to a store and analyze the purchasing behavior of consumers.

In conjunction with the release of this solution, a white paper (available only in Japanese) which summarizes the results of the proof-of-concept (PoC) has been prepared. The white paper contains information on how this solution can affect store operations, in a form of a case study.

TDSL site: http://resources.griddb.com/whitepaper/pdf/whitepaper-sales-forecast-datafluct-griddb.pdf(182Bytes)
DATAFLUCT site: https://datafluct.com/lp/cloudterminal/topics/whitepaper-sales-forecast-datafluct-griddb.pdf(0Bytes)

In addition, an online seminar about machine learning prediction will be held on Wednesday, May 19, where a few key points on how to achieve a more reliable prediction will be explained. Registration for the seminar can be accessed from the following website.
Online Seminar Registration Site: https://datafluct-griddbcloud-0519.peatix.com/

Both companies will continue to provide prediction systems using machine learning in various fields such as power generation prediction and failure detection.

The Toshiba Group aims to grow as a cyber-physical system technology company by combining its strengths in information processing, and digital and AI technologies with the knowledge and achievements in a wide range of business domains that it has cultivated over many years as a manufacturer. Toshiba Digital Solutions Corporation, as a core company in the digital solutions business domain promote digital transformation, and provide system integration and service solutions using AI and IoT. GridDB is a scale-out database developed by Toshiba entirely in-house, drawing on its extensive expertise in diverse industry verticals. GridDB ensures efficient accumulation of massive volumes of time series data, and deliver scale-out performance. GridDB Cloud is a cloud data infrastructure that provides GridDB as a managed service on a public cloud, which facilitates integration with cloud-native applications.

DATAFLUCT is a data science startup studio that creates new businesses from data and creates value for companies and society with its vision "turning data into commerce". DATAFLUCT has a wide range of analytical experience, from satellite imagery to POS data, and can build algorithms across industries, regardless of technology or data. In addition to supporting DX through accumulated knowledge in developing a wide variety of in-house services, DATAFLUCT is also working on new business development that balances business and social contribution, aiming to contribute to the SDGs through data utilization. DATAFLUCT cloud terminal. is a machine learning platform service that allows users to easily, inexpensively, and quickly build models and analyze data using AutoML (automated machine learning) through a simple data upload to the multi-cloud environment such as AWS, Azure, and GCP. Not only that AutoML saves time, it also enables users with minimal knowledge of machine and deep learning to create highly accurate models.

*3　News Release Announced July 13, 2020 (Japanese only)
https://www.global.toshiba/jp/news/corporate/2020/07/pr1302.html

• GridDB is a registered trademark of Toshiba Digital Solutions Corporation in Japan.
• Amazon Web Services, AWS, and the "Powered by Amazon Web Services" logo are trademarks of Amazon.com, Inc. or its affiliates in the United States and other countries.
• Microsoft, Azure, and the "Azure" logo are registered trademarks or trademarks of Microsoft Corporation in the United States and other countries.
• Google Cloud Platform (GCP) and the "Google Cloud Platform (GCP)" logo are registered trademarks or trademarks of Google LLC.