Introduction

Cost pressures, supply chain disruptions, compliance, and changing customer demands require identifying production bottlenecks, detecting quality issues, and predicting machine failures. Spotfire, a visual data science platform, helps solve complex manufacturing problems, reduce costs, improve operations, and increase profitability. Whether you’re a process, yield, or test engineer, fab manager, or in supply chain management, you face daily data challenges.

Recent Engagements

September 2024: Spotfire hosts a TAF Webinar on Visual Data Science for High-Tech Manufacturing. Join Brad Hopper and Alessandro Chimera of the Spotfire Vertical Markets team as they discuss the latest in visual data science capabilities to cover high-value challenges in the industry. Some featured use cases covered in the webinar are process characterization, anomaly detection, and root cause analysis, among many others.

July 2024: The Spotfire team kicks off the TAF Webinar Series, with the first webinar providing an executive overview of visual data science. Follow along as Michael O'Connell, Chief Analytics Officer at Spotfire, discusses visual data science, visualization in Spotfire, recent trends in AI, and industry specific applications. The full webinar, along with additional resources, can be found here.

July 2024: Members from the Spotfire team attended SEMICON West 2024 in San Francisco, a leading semiconductor and microelectronics manufacturing industry conference. At SEMICON West 2024, the use of AI in the manufacturing industry was one of the largest conference themes, with several sessions talking about use cases such as yield optimization with AI, virtual metrology, and more. Check out this Spotfire blog post for more information about our time at SEMICON West 2024!

May 2024: Michael O'Connell, CAO at Spotfire, visited Japan and Korea and gave Spotfire seminars focusing on Visual Data Science and AI innovation in manufacturing. The Japan seminar was hosted by NTT Com Online and the Korea Champions event was hosted by P&D Solutions. The seminars covered a variety of topics, such as how to use Spotfire to optimize yield, how Spotfire Copilot can be in multiple languages, such as Japanese, Chinese, and Korean, and utilizing Spotfire in the process of parametric testing. To learn more, check out this Spotfire blog post.

For more information on recent engagements and upcoming Spotfire events, check out the Spotfire Community Events page.

Spotfire Solutions by Use Case Domain

Products: Quality and Reliability

Spotfire helps manufacturers to identify, understand, and minimize problems due to process variability, incoming supplies, test, or design. An intensified interest in product quality and reliability analysis is being driven by a number of market forces. Products today are more complex, have shorter lifecycles, and are increasingly connected. Reliability failures are more visible and sometimes more costly than ever before. Meanwhile, the forces of technology, globalization, and regulation make our quality and reliability calculations more complex and urgent. Many of the world’s leading manufacturers are turning to Spotfire to identify issues earlier, respond more rapidly and effectively and then build better products.

Wafermap Pattern Recognition and Classification

Identifying patterns of interest in big data is complex, but critical to high-value manufacturing use cases. Explore Spotfire's novel approach to big data pattern recognition as applied to semiconductor wafer maps. Using a combination of machine learning techniques, business users can identify patterns quickly and accurately from a large amount of data. Using these patterns, Spotfire users can train and deploy a model to classify new wafers in real-time.

For more information:

• Watch a full demo of this application
• Try it out yourself in the Spotfire Demo Gallery
• Download it from the Spotfire Community Exchange
• Learn more about it on this Spotfire Community article

Wafer Defect Analysis

Patterns on semiconductor wafers are often complex and hard to understand. With so many measurements made across so many chips, it is arduous to track down how and when a pattern starts to form. This Spotfire application helps you navigate through this challenge. A combination of visual analytics and data science aids root cause analysis by comparing patterns on wafers at different points in the manufacturing process.

For more information:

• Try it out yourself in the Spotfire Demo Gallery
• Learn more about it on this Spotfire Community article

Wafer Scratch Detection

Failing to identify a scratch on a wafer can be costly. The challenge with scratches is that they come in all shapes and sizes. Because of this, typical machine learning models that are trained to identify a certain shape might not be effective. We built this application to address this challenge by using advanced clustering techniques and principal curves to dynamically identify scratches of varying shapes and sizes. 

Finding one scratch can be important, but our application also leads you to find similar patterns, uncovering more scratches. Once a significant amount of scratches are found, we can mathematically understand what defines scratches and automate their detection.

For more information:

• Try it out yourself in the Spotfire Demo Gallery
• Learn more about it on this Spotfire Community article

Wafer Acceptance Test Correlation Analysis

Rigorous wafer testing processes such as wafer acceptance testing often come with large amounts of data, which can be challenging to analyze and make meaning of. This application addresses this challenge by measuring the strength of association, calculated as the Cramer's V measure, between streamed wafer test measurement data and a chosen target variable. In turn, this identifies which sensors and parameters are most correlated with a failed test. With these insights, users will gain direction and guidance on where to prioritize their troubleshooting efforts. Having streamed data also allows these correlations to be monitored in real time, and facilitates quick, data-driven decision making.

For more information:

• Try it out yourself in the Spotfire Demo Gallery
• Learn more about it on this Spotfire Community article

Product Traceability

See how Spotfire and Statistica can help understand the effects of processing on product characteristics using the Product Traceability add-on.

Stability and Shelf Life Analysis

Stability analysis is the study of how drug product potency degrades over time. The primary statistical quantity of interest is the expiration date or shelf life. Typically, a drug product is manufactured in batches. When estimating the shelf life of a medication, it is necessary to evaluate how the batches differ with respect to the potency degradation of the drug product over time. 

You can find an overview of the solution here and more details here.  

Design of Experiments

Design of Experiments is an important tool for experimentally identifying the most important factors and finding their optimum settings in order to improve processes and products. Spotfire and Statistica have comprehensive capabilities for the design and analysis of fractional factorial, Box-Behnken, Central Composite, Optimal, Mixture, Taguchi, and a number of other design types. It also features a prediction profiler for simultaneous optimization of multiple responses. 

Processes: Process Control and Anomaly Detection

Anomaly detection is a step towards resilience that will serve any manufacturer well. Manufacturers with mature anomaly detection capabilities achieve substantial operational cost reductions from reduced defect and scrap rates, improved quality and reliability, prevention of unplanned equipment downtime, and even optimized energy consumption. Typically, most data streams simply confirm normal operations and provide no new actionable information. However, when data shows an anomaly, it can indicate something has changed or is behaving abnormally leading to actionable insights about how to correct any issues before they become widespread or time-consuming. 

For more information on Anomaly Detection with Spotfire, check out this community landing page.

Univariate Process Control

Control charts are widely used in Manufacturing, Energy, Telco, Technology, and many other sectors. They are the foundation of early warning systems that monitor key metrics, detect deviations from the baseline, and generate automated alerts. Spotfire supports many types of Shewhart (univariate) and multivariate charts: integrated limits generation, storage, and deployment, selection of rules to detect out-of-control points, tagging and annotation, management and operations dashboards, periodic or real-time alerts, process capability studies and root cause drill-downs.

Process Control Summary with drill-down to Control Chart

Our downloadable solution for Process Control, Monitoring, and Alerting in Operations applies statistical methods to monitor and reduce the variability of measured processes. It is an easily configurable quality control solution built with Spotfire and Statistica that can monitor large numbers of parameters and produce automated alerts when rules are violated. Data is visualized in linked Spotfire dashboards, and Statistica is the calculation engine supporting rules, alarms, and alerts.

• Read more about the solution here
• Watch a short demo of what the solution does here
• Watch a more complete demo of the solution and its architecture here
• Try a live interactive demo on the Spotfire Demo Gallery
• Download this solution from the Exchange

Multivariate + Comprehensive Process Control Solutions 

Statistica has comprehensive out-of-the-box Process Control capabilities including Quality Control Charts, and the Multivariate Statistical Process Control for automated monitoring of large numbers of charts. The capabilities are tightly integrated with Spotfire, via the Statistica-in-Spotfire data function, to enable calculations in Spotfire Data Science with data visualization in Spotfire.

Using AI to detect complex anomalies in time series data

The Spotfire Data Science team is actively engaged in developing applications of Deep Learning Autoencoders for Anomaly Detection in Manufacturing. In a dynamic manufacturing environment, it may not be adequate to only look for known process problems, but also important to uncover and react to new, previously unseen patterns and problems as they emerge. Univariate and linear multivariate Statistical Process Control methods have traditionally been used in manufacturing to detect anomalies. With increasing equipment, process, and product complexity, multivariate anomalies that also involve significant interactions and nonlinearities may be missed by these more traditional methods. This is a method for identifying complex anomalies using a deep learning autoencoder. Once the anomalies are detected, their fingerprints are generated so they can be classified and clustered, enabling investigation of the causes of the clusters. As new data streams in, it can be scored in real-time to identify new anomalies, assign them to clusters and respond to mitigate potential problems. These tools are no longer the exclusive province of data scientists. After an initial configuration, the method shown can be routinely employed by engineers who do not have deep expertise in data science. 

• Watch a full demo of this Spotfire application.
• Try the application for yourself and interact with the anomaly detection features.
• Download this solution from the Spotfire Community Exchange.
• Visit the Anomaly Detection community page for more information about anomaly detection.

Advanced Process Control

Advanced Process Control is an application of digital twin technology that involves the use of sensor & metrology data to implement real-time tuning and control of processes. This facilitates greater control of process variability than is achieved with the Basic Process Control techniques above. Techniques include feed forward, feed backward, virtual metrology and predictive process modeling.  

Comprehensive Industrial Statistics and Six Sigma

See a complete list and description of all Spotfire Data Science Industrial Statistics and Six Sigma Solutions.

Machines: Predictive Maintenance & Anomaly Detection

The expansion of connected sensor data creates new business opportunities for monitoring machine performance and failures in the field and on the factory floor. Service organizations have up-sell opportunities to offer options to their customers for maximizing the value of their assets. Manufacturers can increase uptime, minimize costs, and optimize processes for expensive equipment on the factory floor. Spotfire helps organizations optimize maintenance schedules by monitoring and responding to key signals in sensor data. In general, fixed assets, vehicles, plants, machinery, communication devices and computers, and even buildings are becoming smarter. But they are also becoming more complex and more costly to repair. Spotfire can help you understand these machines more fully, monitor them in real-time, and react faster to impending issues. Spotfire supports the following maintenance use cases: predictive maintenance with automatic notification of impending failures, minimizing scheduled maintenance costs, and root cause analysis of equipment failures.

Monitoring Machine Sensor data with Spotfire Streaming Analytics - Spotfire accelerators jump-start building end-to-end analytics solutions. See what's new and watch a demo of the Spotfire Intelligent Equipment Accelerator. You'll learn how to capture and analyze IoT sensor data in real-time and integrate using industry-standard protocols like OPC UA, OSI PI, MQTT, and Web Services, or build your own. In addition, how to apply custom validations, cleansing policies, rules, and feature statistics to data feed to identify trends and gain insight, and how to use real-time model execution for anomaly detection and classification.

Factory: Monitoring, Maps, Anomaly Detection & OEE

Modern factories are populated with complex, expensive equipment. Manufacturers want to extract the greatest value from their factory equipment by maximizing equipment uptime, product throughput and quality and minimizing cycle times. Identifying bottlenecks in processing, taking proactive action in response to developing situations, and increasing operational system awareness are all key themes in sensor-driven manufacturing monitoring.

Overall Equipment Effectiveness or OEE is a high-level measure of equipment productivity. The OEE model combines measures of equipment availability, performance and quality. 

• Availability is the percentage of time that the equipment is available to operate, or Uptime. Scheduled downtime, unscheduled downtime and non-scheduled downtime (holidays or training) all contribute to availability losses.
• Performance is the speed at which the Work Center produces product as a percentage of its designed speed. Performance losses are categorized as either due to Rate or Operational inefficiencies. Rate losses are caused by equipment running slower than theoretical speed. Operational Losses may be further broken down into Engineering and Standby Losses. Engineering losses occur when production turns equipment over to engineering, often to perform tests or experiments. Standby losses occur when the equipment is available but there is no product or operator to run it.
• Quality is Good Units produced as a percentage of the Total Units Started. Sometimes referred to as First Pass Yield. Rework and scrap contribute to Quality losses
• OEE is calculated by multiplying Availability, Performance, and Quality percentages together.

To learn more about the OEE and Plant Productivity solution in Spotfire, check out this article, and try out a live interactive demo on the Spotfire demo gallery.

Data Replay Accelerator

The Data Replay Accelerator for is used to capture and replay in real-time data from historian systems like OSI PI, OPC UA, or a other repository. When an anomaly or trigger condition is detected, it retrieves a configurable amount of data before and after the anomaly, and gives operators the capability to replay this data for observation in a Spotfire dashboard.

Download the Data Replay Accelerator from the Exchange, and check out this community article for more information on the accelerator.

Data Historian Accelerator

The Data Historian Accelerator captures real-time telemetry from data historians like OPC UA and OSI PI. A custom HTML5 web interface provides the user the ability to visualize the object hierarchy of the historian, and create subscriptions on the nodes or tags of interest. The accelerator receives this data in real-time and assembles the points into logical data sets that can then be passed to business rule modules that implement decision tables or data science models.

Download the Data Historian Accelerator from the Exchange, and check out this community article for more information on the accelerator.

Intelligent Equipment Accelerator

The Intelligent Equipment Accelerator contains components to allow monitoring of production line performance against established metrics using Overall Equipment Effectiveness (OEE). It captures data feeds from sensors on production equipment, validates the feeds, and evaluates the data against configurable business rules. It includes components to visualize all these activities in a custom web dashboard, allowing operators to take corrective action when production issues are identified.

Download the Intelligent Equipment Accelerator from the Exchange, and check out this community article for more information on the accelerator.

Supply Chain: Demand and Transportation Logistics

Recent AI, automation, and data management breakthroughs help supply chains of all types sense and respond to real-time conditions, like your body's nervous system. They can sense demand, operations, and volatile conditions to respond to what's happening now.

Continuous Supply Chain - Continuous Inventory Tracking

Building a resilient supply chain requires connecting all your data wherever it is, unifying it to achieve consistency, and applying AI to develop deep insights for decision-making and automating manual processes. This accelerator features real-time inventory tracking, with real-time stock alerts, store deliveries, distribution center alerts, and transport logistics optimization.

Download the Continuous Supply Chain Accelerator from the Spotfire Exchange.

Other Use Cases

Below are a list of additional use cases and solutions possible with Spotfire. For more information on any of the following use cases, please contact datascience@spotfire.com.

• Equipment Commonality Analysis
• Reliability and Warranty Analysis
  • Monitor and predict claim rates, analyze root causes of reliability failures, and analyze warranty repair and call center activity.
• Machine Learning for Root Cause Analysis
  • Apply machine learning techniques to improve yield and quality, perform predictive and condition-based maintenance, micro-segmentation of markets, and resource optimization. 
• Big Data Product Digital Twins
  • Perform real-time, continuous analysis of manufacturing equipment sensors and process data with digital twin technology. 
  • For more information on digital twins, check out this Spotfire page.
• Six Sigma Connected Production Platform from Genware/DataShack
  • The Connected Production Platform provides manufacturers and producers with Spotfire Hyper-Converged Analytics required for a Next Generation Intelligent Digital Plant, along with a Blueprint that processes source data, applies predefined machine learning models, and includes advanced analytics, all in support of predicting. Learn how this platform provides access to data to monitor live performance across the entire plant value stream, utilizing the Six Sigma Methodology.
  • For more information, check out this video demonstration.
• Analytics for Formula 1 Racing with Mercedes AMG Petronas racing

Customer Success Stories

Western Digital

Ahmer Srivistava from Western Digital shares how the use of Spotfire analytics has transformed high-tech component manufacturing to meet growing demand.

Ahmer's keynote segment starts at 30:00 in the video.

Over the last 8 years, Spotfire analytics has become the standard platform used by Western Digital engineering and operations to view and analyze manufacturing operational data. At wafer factories in Silicon Valley, the use of Spotfire software as the data analytics and visualization standard for factory operations has grown considerably. We will look at this integration in terms of architecture, data infrastructure, user groups, and business processes over three time periods and showcase solutions that made it possible to significantly increase efficiencies in the way we work. These use cases cover yield analytics, metrology, sensor data, and operational metrics described from the perspective of purpose, implementation, benefits, and differences with and without Spotfire software.

Pfizer

In this session you will learn about Pfizer's pathbreaking efforts to become an AI-driven organization, utilizing an innovative Manufacturing Intelligence Workbench concept, designed with the goal of driving IT/OT convergence across the interconnected network of manufacturing sites, providing real-time access to data from all sources, and enabling scalable and accelerated AI/ML deployments in support of manufacturing operations. The MI Workbench enables Pfizer manufacturing sites to achieve Digital Plant Maturity and the Factory of the Future vision by becoming more predictive and adaptive and empowering the shop floor. This cloud-based platform provides high-performance environments to develop and deploy a myriad of analytics capabilities, including advanced web-based dashboards and reporting tools, real-time multivariate monitoring and control, golden batch analysis, digital twins and soft sensors, and AI/ML-based predictive models. 

Hemlock Semiconductor

Successful analytics deployment requires focusing not just on data and tools, but also on the people that use them. Leveraging Hemlock Semiconductor's successful deployment of Spotfire and Data Virtualization software, this session discusses strategies for building an analytics culture and overcoming challenges along the way. Topics include centralized vs. decentralized approaches, leveraging early adopters, benefits of unpolished data, adapting skill development to meet users where they are, and modifying your approach as the enterprise matures.

Texas Instruments

Texas Instruments uses Spotfire for advanced analytics across the company, from manufacturing to sales. Enabling stakeholders to aggregate and analyze vast quantities of data. This session will highlight two examples of the creative delivery of advanced and actionable insights to TI's sales and pricing organizations. We will dive into the tool architecture and how the combination of scripts, data modeling, procedures, and tagging of a visualization drives specific actions across TI.

Keysight Technologies

Designing the flow of data from complex datasets into actionable information in the form of fast-loading Spotfire analysis requires a combination of techniques. Through a series of case studies, this session will highlight some successful strategies involving Python scripting, Spotfire Automation Services, advanced information link design, web caching, and data virtualization.

Semiconductor Customer & Partner Success Stories

• STMicroelectronics
• Hemlock Semiconductor
• Keysight
• More Spotfire Customer Success Stories

Resources

• Spotfire for Manufacturing
• Anomaly Detection in Spotfire
• Spotfire Demo Gallery
• Download Manufacturing Industry templates, data functions, and accelerators from the Spotfire Exchange
• Spotfire Training
  • This is the starting point for an extensive, constantly expanding collection of linked article pages covering Spotfire Analytics capabilities.
• Spotfire Resource Library