What Happens When You Don’t Trust The Data?
Picture source: http://www.mobgenic.com
BY PAUL TEMPLIN
INDUSTRIAL ENGINEERING – MAY 2013; VOLUME 45; NUMBER 5
How the world has changed? Or has it? We now live in data rich environment. Today, each widget comes off the production line with its own “birth certificate” of rich parametric data. Rolling down the line, process steps are recorded in real time, and test data is linked to each widget. Yet today, more than ever, we question the data.
In the past, we were starved for data and each morsel was precious. We understood the limitations of the data and used linear assumptions to massage the data and fill in gaps. With increasing sophisticated, we query our databases. While in the past, we had to type data into spreadsheets or calculate metrics with a pocket calculator, today’s database tools allow us to query multiple tables in real time and automatically port data into spreadsheets or statistical tools to perform sophisticated data manipulation. Ten of thousands of records can be analyzed and summarized; yet, when we do so, confusion often results as we see inconsistencies. Corresponding data in various tables does not match or is wrong. Tenet of database design is never to store the same data in multiple tables or databases. Nowadays, with increasing velocity, data sets are passed from one type of database to another using translation matrices and other sophisticated types of data translation tools. With each translation of the data, data congruence degrades. Somehow, data is lost during the translation. If this was our only challenge, it would be bad enough. But as data congruence degrades so does our confidence in the data.
Data still needs to be treated like a precious resource. Industrial engineering rigor to be applied to data collection, identifying what data collect and every other step in the process. Apply other lean principles to data collection and storage. Avoid the pitfall of the data hoarder. Do not end up with a garage full of confusing and contradictory data. Trust the data you have. If you don’t, fix it, do not duplicate it.
This sounds easy, but of course it is not. Creating a database of meaningful and accurate data is hard work, but having data you trust and believe is priceless.
The full version of the article is available in IIE Laboratory. It is also readable online for IIE member through accessing the iienet.org website. Contact Maya (President of IIE BINUS University Chapter) at firstname.lastname@example.org for more information on the IIE membership.