The dirty data costs the US economy. UU Up to $ 3.1 trillion a year, and organizations have tried to deal with them using methods such as deduplication, standardization, or even manual deletion or correction of data. broken or incomplete.
There are tools that do, but there are times when manual data correction is required. Needless to say, companies are drowning data and are unable to assert themselves that data can not be cleared quickly enough.
SEE: Sensor'd Enterprise: IoT, ML and Big Data (ZDNet) | Download as PDF (TechRepublic)
In summary, it's time to rethink data cleansing to be at the top.
How are you, so you can continue?
1. Redefine data cleansing as a problem of commercial integration
A big problem concerns data that does not work with other data because it is in a different format or is known by a different name. It is possible that the same data in different systems may have different field sizes and be named differently from department to department.
Problems like these are more important than gaps and data inaccuracies. They might reflect different systems that duplicate each other because the different business units do the same things and do not even recognize them because they use different overlapping systems. In this case, it is best to meet all users and departments to decide which system to remove and which system to keep to avoid confusion. You may have a one-time job to transfer data from a previous system. System to a new one, but you will never have to do this work again.
2. Use tools and services that minimize the impact on business processes
"When companies come to us, they tell us that they are scattered across many different systems and formats, and that the quality of that data is low," said Vivek Joshi, CEO of Ettyle , who helps manufacturers. , Sell more products for your after-sales customers. "Businesses want to understand this data and do not know where to start, and they involve us because we have the experience of the algorithms and analytics we use to cleanse and align that data so that it can be used."
Joshi says his company also makes data integration less painful for customers. "To do this, we add a click tab to existing system menus so users can just click on and access the data and analysis, then return to the system they're working on," he said. he explained. Integrating data analysis into a given system platform can be so easy. "
The advantage of companies choosing tools and methods from external suppliers is that the data is readily available and the penalty for verifying the business process is virtually eliminated.
SEE: IT Guide for Big Data Security (Tech Pro Research)
3. If necessary, avoid overhauling business processes
While rapid data and systems integration tools are valued, organizations must always understand the need to review business processes so that the organization is operationally aligned with its mission and customers.
For example, a product engineer who has once abandoned a design for manufacturing and has forgotten, can continue to collaborate with manufacturing and even customer service as part of reducing product cycles. If the engineering does not remain "connected," you may lose valuable information about how to improve products that are built on customer feedback for product performance and ease of use.
4. Decide what data you want to keep
It is not possible for companies to cleanse and store all the data flowing into their business, even with the most effective data preparation and cleaning tools. That's why it's so important for the IT department to work with the different departments to determine what data is stored and what data is backed up. Once you've made the decision, your concerns about the critical data persist and your data is kept clean, it's easier
See also:
Picture: iStock / SvetaZi
Aucun commentaire:
Enregistrer un commentaire