This case study explains how to obtain a realistic assessment of the improvement potential of a manufacturing process, in the pharmaceutical industry.
- Data Science
- Time Series Analysis
- Predictive Quality
- Dashboards
- Data-driven assessment
Key challenges
Our client is a pharmaceutical industry, world leader in the management of pain and related diseases. They are producing drug solution and monitor each step of the production. In this context, they had collected and stored a lot of data. They thought they could improve their production yield by up to 5% by perfecting their production process. In order to confirm their intuition and be able to improve their process, our client was looking to:
- Get a realistic assessment of the potential for improvement of the manufacturing process.
- Identify the critical processes parameters that influence their product quality.
- Be able to analyse up to 10K parameters (considering that not a single analytics tool was implemented yet).
Our approach
In order to meet our client’s challenges, we used a RaaS (Result-as-a-Service) approach: our client provided us with its data and we integrated it directly into our own SAS-based analytics platform.
We started by reviewing the data and assessing the potential by identifying and correcting data quality problems. Then in order to understand the context and how the different parameters of a batch influence the process, we conducted interviews and workshops with the business teams (including engineers, managers, etc.).
Followed, we focused on the analysis by transforming all historical data into comparable analysis format. We did a deep assessment of the data and thus reduced irrelevant and redundant process parameters to essential influences. Indeed, by using advanced analytics methods, we were able to reduce the number of parameters to be analysed from 10K to just a few ones (around 50 in the end).
We were able to reduce the number of parameters to be analysed from 10K to just a few ones (around 50 in the end).
Nicola Zäh, Industry Analytics Team Manager
We finally created a dashboard using SAS in order to characterize and evaluate significant influences, and discuss the results with our client’s teams.
Benefits
- Ability to actually visualize the data: Data-driven decisions vs Data-Inspired
When we started this project, our client thought he would improve his process up to 5%. However, after a honest analysis (data-driven assessment) of the potential, the results showed that there was at most 0,5 – 1% real room for improvement.
- Control of costs and time through minimal investment
Through the RaaS approach and the data integration into our own SAS-based analytics platform we enabled our client to get quick data-driven results without investing in a complete infrastructure and implementing analytics tools.
- Actionable dashboard and concrete recommendations
We delivered an actionable dashboard with recommendations to our client. Then they can prioritize which element have to be improve first. We also identified and eliminated major data quality issues they had in the data collection process (with various sources systems and structures).
- Smart combination of the engineering and the data points of view
By combining the data perspective to the business one, our client gained new insights and perspectives on their machine. For example, it helped a lot from an engineering point of view to know where to put the emphasis on if you want to inspect a machine.
Technologies & Partners
![SAS and Python technologies Data & Analyatics use cases](https://positivethinking.tech/wp-content/uploads/2021/12/Template_Use-Cases_mining-e1638861243134-edited.png)
![Newsletter Subscription](https://no-cache.hubspot.com/cta/default/7985071/9c9a1662-e108-4e01-bf0f-1f9808b0d53e.png)