Technology

Data Analysis to Accelerate Innovation
A Highly Accurate Data Analytics Tool Refined
Through Sony's Diverse Businesses

  • Shingo Takamatsu

    R&D Center,
    Sony Corporation

  • Hideki Yakura

    Sony Assurance Inc.

Since entering the 2010s, we have seen an acceleration in the movement to aggregate a flood of data in our daily lives, analyze it, and use it for business or lifestyle purposes. There is a particular focus on data analytics, whereby future results are predicted based on past results using statistical algorithms and machine learning. We spoke to Shingo Takamatsu, R&D Center, Sony Corporation, who has been involved with this research. “Data analytics enables companies to take more effective measures by giving them a better understanding of customer purchasing behavior or helping them to analyze manufacturing processes to improve yield based on manufacturing information,” explains Takamatsu. “However, while data collection has become more commonplace, analyzing and utilizing vast amounts of data of various types requires a high level of expertise.”

What is data analytics?

A technology that predicts future results based on past data through machine learning.

Example of a recurring contract

Takamatsu and his team have been developing a data analytics tool called “Prediction One*.” He says that a distinctive feature of this tool is that even someone who isn't a data analytics expert can use it with ease. “We've created algorithms to handle various types of data that our users might want to analyze. Users can perform data analyses with a single click, as the tool automatically selects the appropriate analytical method for the input data. The tool works as a desktop application, and this eliminates the need to send data outside the company.”

*Prediction One
A data analytics tool featuring a simple and intuitive UI that can be easily operated without any expert knowledge. Developed using Sony's proprietary technology, this tool is user-friendly, accurate, and provides reasoning for its predictions. Additionally, it can operate on ordinary systems such as a laptop, meaning that no special equipment is required. This service was launched to the public in Japan from Sony Network Communications Inc. in June 2019.

Takamatsu recalls that he is able to develop these features due to Sony's diverse businesses. “When we introduced the data pre-processing and model construction technology developed by R&D, we received many requests for data analysis from within the Sony Group. As we assisted those businesses, we also developed tools to meet various data and analysis needs. Based on the rich data samples we received, we improved our algorithms and eventually created a tool capable of predicting many cases in advance. And since the tool was widely available within the Group, we had many test users right from the prototype stage. Based on the feedback from all those users, we identified new issues, which ultimately allowed us to improve the tool.”

Hideki Yakura from Sony Assurance Inc., who has used the tool since the beginning of development, commented, “They listened carefully to all of our feedback, so the tool became very user-friendly. Naturally, it's also very accurate and allows one to get a good grasp of customer trends. Furthermore, it also displays its reasoning, making it easier to explain improvement proposals based on forecasts. At the moment, we're working on our usage skills, namely, ‘What to analyze and how to use the results.' Together with Takamatsu, we're searching for even more effective usage methods.”

“We believe that the successful adoption of data analytics will benefit both the company and its customers,” explains Takamatsu. “If this tool can accelerate progress in society, then surely it's worthy of being called ‘innovation.' Moving forward, we would like to collaborate internally to further improve this tool, while also making it available to people outside of the Group.”

Stories

Page Top