Friday, November 27, 2009

Databases

Some of my earlier posts have had to do with information (or data) visualization. Before you can visualize data, it must be readily available. In today's age of computers, your data is certainly electronic but if it is spread across multiple servers, directories, and files it is unmanageable and not available. This is where relational databases come in.

If your data is in a well-designed relational database, you can access it in a variety of ways. Most data analysis software (e.g. Excel, MiniTab, Origin, Quality Analyst are a few that I've used) has wizards that make it easy to develop SQL queries to retrieve data. With a minimal knowledge of SQL you can modify the wizard queries to make them even more powerful. The important part is to have your data in a database to begin with. While a LIMS or another database may seem like an expensive investment, it will pay off in the long run by giving you many opportunities to examine your data and find answers to your problems in data you have already gathered rather than having to design new experiments.

In today's age of tight budgets, many companies want concrete justification for purchase and use of a database. The problem is that a database will help you solve problems that you haven't even imagined yet and so it is hard to provide evidence of the "hard" savings that a database will provide.

Anybody with an eye towards the future (not just the next 3 months) should have a relational database for storing their data and begin inserting data immediately. Without data in your database, it will be useless.

No comments: