Industrial processes: defining your data analysis strategy

Industrial processes: defining your data analysis strategy

Data analysis strategy: set your objectives

It is both common and legitimate to say: “I have loads of data but what can I use it for?” Yet it can be counterproductive to try to understand data without establishing a framework and identifying potential actions. Setting objectives is fundamental to a strong data analysis strategy. Start by asking the following questions:

What are my objectives?

These may include:

  • knowing more about my process to reduce unpredictable quality;
  • reducing energy consumption per ton of production uncorrelated with volume;
  • improving predictions about the capacity of my process for the recipes used.

How can improvement be measurerd objectively?
How do I measure my current performance?

Note that the more precise and measurable the objective, the easier it is to roll out your data analysis project.

Align your data analysis strategy with your organization processes

Identify key roles

Like any other process, your data analysis strategy and process depend on the key people in your business. You can rely on the following members of your team to enable process improvement:

  • the operational team: workshop foreman, team leaders, operators, technicians, etc.;
  • the support team: process engineer, continuous improvement engineer, etc.;
  • and if your team includes them: data analysis/statistics experts.

Each of these key staff members have specific expertise to contribute to your analysis. It is important they have a defined role in the project.

Foster collaboration

The best recipe for effectively achieving objectives is to combine expertise in processes, operations, and statistics. Such expertise is rarely found in one person or even in one team, and maybe not in-house. It is therefore important to foster collaboration for your data analysis strategy.

Data analysis analysis: make way for action and controlled experiments

The solutions to your problems are rarely found in the analysis and interpretation of historical data. They do, however, often feature robust solutions for improvement and allow you to:

  • Undertake actions to gain more control over critical process parameters.
  • Develop plans for controlled experiments to measure concrete progress. It can be useful to begin by establishing that data analysis supports controlled experimentation to achieve improvement for the issues identified. Confirm that time will not be lost doing countless tests without exploiting the results.
  • Design new technical solutions to achieve your goals.