Impact of Intelligent Data Management Systems
For the sake of simplicity, we say that intelligent quality control systems impact the optimization of three fundamental dimensions: data capture, analysis, and sharing.
However, and although it is not expressly mentioned among the advantages, there is a preliminary phase that is the basis for the success of any data mining tool: defining as accurately as possible what data we need, what they can provide us with, and what we intend to obtain with them.
Data definition
Indeed, the definition and planning phase is the basis for the success of a data processing project. Before we start capturing data, we must answer a series of questions of the following type:
- Where does the data originate and how often?
- Which data sources have the potential to provide added value?
- How do we guarantee its reliability/homogeneity/veracity?
- How are we going to store them?
- Is it necessary to analyze them in real time?
- Can we combine our data with other external data to help us look for valuable correlations?
- How useful do we expect this data to be, and what do we want it for?
Data models
Since the answer to these questions will shape the model to be implemented, it is very important to be able to answer them adequately, because it is in this phase where a large part of the success or failure of a data management project is gestated. Let’s keep in mind that the ultimate goal is to be able to have adequate information for efficient decision making, so if the model is wrongSo if the model is poorly defined, decisions will be uncertain to say the least.
“Let’s keep in mind that the ultimate goal is to be able to have adequate information for efficient decision making, so if the model is poorly defined, decisions will be uncertain to say the least.”
For example, if quality control data are expected to come from both industrial processes and laboratory tests, it is very important to define how they will be correlated (through the manufacturing batch, for example) so that the system can interweave the information and detect possible changes in behavior patterns, deciding whether it is an anomaly or not; in this way the system will continuously learn.
Or, in the case of references that form a subassembly, it will be necessary to foresee the need to have a list of components to maintain traceability and ensure the integrity of the information.
It is about not getting lost in the data, and avoiding that a Big Data project becomes one of Big Chaos. We must always think about solutions that add value rather than focusing on amassing figures and data; as the Pirelli slogan said, “power without control is useless”, and having hundreds of billions of data without knowing what to do with them or, even worse, having the wrong collection and analysis strategy, is really a waste of time: we will have a graveyard of data instead of useful information for decision making.
