HANDLING INSTRUCTIONS for gathering information from Big Data

Our approach

Watch over diversified consumer activities

Point Three provides a total solution for various aspects of business: from business planning to administrative reform promotion, IT deployment, and assessment.

<Our view point> Data Source and Information

1. Basement: Flat information
he first step is to identify the basement, in which information is arranged in terms of TIMING and COMBINATION.
TIMING means “to collect information to USE immediately.” Information may be new or old. Even if information was not accessed recently, it is extremely valuable to know about it at the time of access. The next step is to consider “credibility of data” in relation to Big Data, given that useless information could be included in Big Data. However, determining the credibility of data is not as simple as verifying if data are correct. Credibility can be seen after COMBINING different data. The basement may then be ready after collecting timely and trustworthy information.
2. Value adding: From many angles
The value of information is limited into itself even if combined with data. Information should be taken into consideration with the image of additional meaning, because Big Data are changing and growing rapidly. In this Big Data age, setting an assumption at an early stage could be dangerous, as sample data may be limited. The point of using Big data is to look at data from many angles until the information, when combined with related data, can be adjudged as roughly correct or definitely wrong. If it seems roughly correct, then such information should be analyzed before assumptions are set.
3. Practical use and accumulation of data: Building a body of information
A body of information can be built after the above processes 1 and 2 are completed. Big Data can contain a wealth of information, and we can find correlations among various pieces of information. Hence, we can expect another cycle of information utilization in viewing data from many angles and piling the blocks of information.


How to set the business parameters to meet the future demand for EV cars
“How much is the actual selling price of lithium-ion battery for automobiles?”

[ Background of the problem ]

Determining the projection of the electric car demand is the most important indicator for the production of lithium-ion battery.

  • ・Wide range in the researcher forecast: One to four million cars for year 2020.
  • ・No researcher has explained the difference between their forecast and actual data.
  • ・No report shows how many electric cars are actually sold.

[ Approach ]


[ Image of our assumption and verification ]

Data Gathering/Analysis
  • ・Collect researchers’ demand forecast for EV cars
  • ・Analyze the gathered information to examine the accuracy and identify the reason for failure
  • ・Analyze the relevance between the sales record and each parameter
  • ・Reinvestigate the sales record and its relevance
  • ・Set assumptions both for advanced and developing countries
  • - Advanced country’s model: Coefficient of appeal
  • - Developing country’s model: Coefficient of environment
  • ・Verify the assumption for advanced and developing countries
  • ・Additional investigation on the social background, traffic situation, and politics in China

[ Efficiency and output ]

These two issues have become clear. Hence, the blind spot of existing reports for the future demand and accurate forecast of EV car sales can be determined.
Simulation of the sales of EV cars for developing countries
Find the factors needed in selling EV cars by identifying the reasons for reluctant sales with the result of factual analysis.

Note: indicators studied for parameters
GDP, population, CPI, oil price, gasoline price, number of gas stations, import tax, tourist income, CO2 emission volume,
health expenditure, number of diesel cars, etc.

Big Dataの活用


  • ・Optimization of resource allocation in the activities of advertising
  • ・Measurement of effectiveness in the advertisement of a product
  • ・Method for assessment utilizing qualitative data

[ Background of the problem ]

Business has become drastically competitive because of several changes, such as the multipolarity of the market, variety of consumers’ taste, and diversification of media. However, this client still depended highly on TVCF for the advertisement of products. The client feels the need to know the primary factor affecting the sales volume and to see if the high cost TVCF is worthwhile, which should be proved by the objective analysis based on data.

  • ・The client did not have useful data to analyze the cost allocation by product and between various media.
  • ・The client lacked experience to review the marketing strategy based on cost allocation and consumers’ buying processes.
  • ・The client could not identify the reasons for the gap between the sales record and its projection with their method of analysis because the body of data was unexpectedly big.

[ Approach ]

Viwpoint Data saurce

[Result of the project]

This client could find their own measurement of effectiveness and the optimal way to use the score, that is, the output from the analysis of both the numeric and nonstructured data quickly.

  • ・To see each of their competitors’ approach and its effectiveness, as well as its impact on the client’s sales
  • ・To classify consumers into groups based on trends, changes, and new signs
  • ・To identify points for improvement in their marketing and promotion after the measurement of effectiveness of the same
  • ・To maximize ROI (Best allocation of investment)

Measurement of effectiveness, utilizing nonstructured data on their competitors

1. Collection and numeric conversion of nonstructured data on consumers
  • ・To develop the logic of numeric conversion from nonstructured data
  • ・To draw two scores from the Internet, such as blog and rating service websites: (1) “response to advertisement” and (2) “appeal of a product”
  • ・Reactivity coefficient: Degree of public discourse on the product and media promotion
  • (Weight-1 X volume of discourse about the measure + Weight-2 X volume of discourse about the message)
  • ・Attractiveness coefficient: Degree of attractiveness based on evaluation axes, each of which affects product perception.
    (Total of the scores by factor of attractiveness, with the formula of Weight X positive reaction ratio)
2. Determine the relevance between the nonstructured and numeric data
  • ・To show the coefficients of reactivity and attractiveness of the product, of both the client and its competitors, along with numeric data, such as sales and market share
  • ・To analyze the impact of competition between other products and that of the client
  • ・To analyze the effectiveness of the measure of the client and its competitors by listing the factors of the nonstructured and numeric data
  • Time series data analysis
  • Competision analysis
  • Relativity analysis

To see the changes and to find the relativity of the factors.

[ Sample output : Relativity between the reaction and market share ]

Sample output