Manufacturer Makes Sense of Quality with Ople.AI

Case Study

Fortune 50 Manufacturer

Market Cap
$30+ Billion

$85+ Billion

Company Description
As one of the world’s largest manufacturers of chemicals and science-based products, the company specializes in Science and engineering, Applications development, Specialized materials and ingredients, Electronics and Imaging, and Nutrition and Bioscience.

“The Ople Platform helped us to quickly develop a more accurate model for predicting the quality of our plastic sheeting manufacturing. The ability to bring in all of our sensor data in near real-time provides the insights we need to not only make decisions on a single run, but also improve quality overall in the long term.”

Key Achievements

  • Improved the accuracy of the predictive model used to manage the quality of their manufacturing process for a plastic film by xx%
  • By automating the development of the new model, the Ople Platform delivered a more accurate model in a fraction of the time of their traditional method
  • Ople’s model explainability and transparency delivered insights into how the model worked, providing confidence to management when deploying
  • The overall quality of the company’s plastic sheeting improved since they were able to detect poor quality early enough in the process to shut down the lines
  • Higher quality products directly translate into higher revenues and profit margins for the company

The Challenges

A large manufacturer of specialty plastic sheeting was looking to improve visibility into their manufacturing processes. In this industry, the plastic is graded and the different grades of plastic – Grade A, Grade B, Grade C – sell at different price points. After every manufacturing run, a piece of the plastic is cut out to test the quality. The main objective is to produce the highest grade plastic as often as possible to maximize revenue and reduce waste.


Their current production line has 120 sensors that monitor and measure the manufacturing process from start to finish – which takes over two hours. The sensors output four different quality measurements at 30-second intervals throughout each manufacturing run. Using these measurements, the company’s data science team created a machine learning model to try to predict the quality of the plastic while it was being manufactured.

Although the model was effective in helping maximize the production process, the company believed a more sophisticated predictive model would help further improve quality and reduce waste. The original model was built by hand, and used a simple linear regression algorithm that the data scientists believed could be replaced with something more sophisticated. 

Initially, the company’s data science team wanted to update the model, but because of a number of higher-priority challenges, they were unable to set aside the personnel and time to fully address the new manufacturing model. Their current process for creating new models was very time-and resource-intensive, involving a number of data scientists and manufacturing personnel to provide the expertise and insights needed to develop the most effective solution.


…the company felt comfortable that their business analysts and business users could quickly become productive with Ople.

The Work

One of the data scientists on the team had been researching automated machine learning (AutoML) solutions to analyze the impact they might have on their team’s productivity. After narrowing the list of possible AutoML vendors, the company selected the Ople Platform for its automation and optimization of the entire data science process, from data entry to deployment. In addition, with its focus on ease-of-use, the company felt comfortable that their business analysts and business users could quickly become productive with Ople. 


The company created a cross-functional team of data scientists and the manufacturing managers in charge of the quality process for the plastic sheeting. This team was trained on the Ople Platform to get started. In addition, as the project progressed, the Ople team assisted in getting the signal processing data from the 120 sensors integrated into the system for real-time insights into the process. 


By including the business perspective from the beginning, the team was able to quickly gather the appropriate data with an understanding of how the data drives insights into the quality of the plastic sheeting being manufactured. An important step in developing a successful artificial intelligence (AI) project, focusing on the business impact – in this case manufacturing quality – helps drive AI success while focusing the team on the goal. 


After gathering the historical data about plastic quality, including all of the accompanying sensor data, the team brought this new dataset into the Ople Platform and let it work through the process of developing the most appropriate machine learning model for predicting quality.

An important step in developing a successful artificial intelligence (AI) project, focusing on the business impact – in this case manufacturing quality – helps drive AI success while focusing the team on the goal.


The Results

Within a few short weeks, the company’s team created a new model that more accurately predicted the quality of the plastic sheeting. Using all of the sensor data as it came in, the company was able to determine in near real-time the quality of the plastic being manufactured in a particular run. This allowed them to check the quality early in the manufacturing run, and if the quality was unacceptable, shut down to adjust the equipment to start a higher quality run. 


Not only does the new model help to improve the quality of the plastic sheeting that is being manufactured while reducing waste, the information gathered and insights gained by the new model are being used to improve the process. By understanding some of the signals that indicate a poor quality run, the company is able to adjust their machines before a run in order to deliver the highest-quality product each and every time. 


With the success of this project, the company is looking to integrate the Ople Platform further into their data science organization, and ultimately push the usage to other groups within the company. By empowering more users within the organization with its ease-of-use and end-to-end focus, Ople is helping to democratize data science within this company. 


With this project, the company learned how easy it is for business users to get up-and-running on the Ople Platform with minimal guidance from its data science team. Currently, they are integrating Ople into more business units across the company to drive the power of AI into more functional areas to enhance productivity, optimize operations, and drive true transformation.

The Ople.AI Platform

The easiest way to accelerate decision making and reduce risk with predictive analytics

With the Ople.AI Platform, you are only a few clicks away from building predictive models to derive optimal business recommendations with reduced risk.