SAS Works with Leading Manufacturer to Develop Visual Data Analytics Programme
Technology manufacturing company uses SAS® Visual Analytics to streamline its manufacturing testing and deliver big data insights to senior management.
A leading technology manufacturer's testing team tests components manufactured by the company for inclusion in its storage area network products. The team engaged in a project to prove to senior management that the analysis of machine-generated log files from the test process would provide greater insight into the overall manufacturing process, and that they should therefore be included in a data warehouse. The team also wanted to determine whether the log files generated during testing could indicate that a specific component was likely to fail in the field, and therefore incur a post-sales cost. Specifically, the team was trying to find patterns or correlations for certain variables within the logs that would indicate a component was likely to fail.
The test harness under investigation heated and cooled each component to defined temperature values, and ran a series of tests. The team wanted to determine whether there were temperature values at which the component never ever failed. Tests at these temperatures could then be eliminated and testing could be concentrated on more vulnerable areas. This process would reduce the length of the overall testing cycle and optimise the manufacturing process.
The team was keen to evaluate the insight garnered from this initial investigation, in order to identify if this approach could be applied across different business areas.
The service the SAS team provided has saved significant time and resources previously dedicated to the testing process. We can process much more data in far less time - using SAS, the time taken to identify variables of real interest has been reduced from months to just hours. We are also excited by the potential of SAS Visual Analytics, particularly for identifying components likely to fail in the field. Preventing component failure will save approximately €20,000,000 per year in reactive engineering costs across Europe.
Senior manager within the company’s testing team
Unyielding Legacy Approach
The company's methodology to date had been to take log files and use a series of C# routines to parse them by variable, and then load them into an SQL Server-based Data Warehouse. Once the log files were loaded into the Data Warehouse, the team used complex SQL statements to extract the information and identify any patterns and correlations. Repeated SQL queries were becoming increasingly unwieldy. As a result, the system didn't deliver maximum value because the SQL queries took a long time to maintain and manage.
"The company's traditional routines failed to deliver the desired result, and the data analysis took months to complete. The programmes used to represent the vast quantities of data graphically were inefficient, and the SQL statements were often terminating prior to completion due to their size and complexity. Using SAS Visual Analytics, the company can explore the data simply and visually, and can now create reports that bring findings to life," said Alex Manek, Principal Pre-Sales Consultant at SAS.
Close collaboration delivers easy-to-digest, highly visual insights for senior management
The manufacturing team provided the local SAS team with a 10GB log file, which was converted into a SAS dataset of approximately 20GB and then loaded into the Visual Analytics infrastructure. The team travelled to the SAS offices in Dublin to undertake the data exploration and produce the reports.
The team produced some interesting results. Indicators from a number of the variables at certain temperatures provided insight into the overall performance of the component at certain stress points. The team was also able to identify certain temperature ranges at which no results were ever recorded. This meant that the calibration of the test harness could be optimised.
From its analysis, the team was able to produce line graphs and histogram reports which outlined the results to senior management. The team was also able to record some screen videos to outline the ease with which these results had been found. In particular, they focused on the Explorer and Reporter functions of SAS® Visual Analytics. The team used the Visual Explorer interface to explore the data and extract variables of interest. This process was followed by use of the Reporter interface to transform the data into graphs which could be presented to management.
Gathering operational insights from data
Applying SAS Visual Analytics yielded data insights not previously obtained from traditional methods. Following the success of the implementation, the team was able to identify other areas where such treatment of data could potentially reduce field failures for components, resulting in a direct saving in support costs. The analytics process was faster thanks to the elimination of superfluous data, and the team could therefore focus analysis on the most relevant and important data. This resulted in significant cost savings.
"The service the SAS team provided has saved significant time and resources previously dedicated to the testing process. We can process much more data in far less time - using SAS, the time taken to identify variables of real interest has been reduced from months to just hours. We are also excited by the potential of SAS Visual Analytics, particularly for identifying components likely to fail in the field. Preventing component failure will save approximately €20,000,000 per year in reactive engineering costs across Europe," said a senior manager within the company's testing team.
Needed to eliminate superfluous tests from the manufacturing test process; needed to present insights from data in a visual format to senior management.
SAS® Visual Analytics and SAS® Data Integration Studio
- Able to process more exhaustive quantities of data to gain better insights
- Increased efficiency - time taken to identify variables of interest is reduced from months to hours
- Cost savings of up to €20,000,000 per year in reactive engineering costs across Europe
- Tool becomes proactive, as well as reactive