Some of you might be squinting your eyes at the screen right about now, scratching your head, trying to figure out how these four things are related. Perhaps even more curious is the image choice for this post. Agreed. A giant robot machine monster is not necessarily something one associates with machine learning.
Well, guess what? Like many cutting edge technologies out there today, Octopai employs machine learning in its unique algorithms. Machine learning enables Octopai to define the dependencies between the various information systems to which it connects in order to be able to create a centralized metadata repository.
How Octopai paves the way toward GDPR compliance
As organizations frantically prepare for GDPR, many are coming to terms with the burning need to automate the way we track the data movement process. They are fed up with spending so much time manually tracing their data’s journey through multiple systems and are looking for a better, more effective, more accurate way to do this. BI groups have to locate and make accessible all the privacy sensitive information they retain about EU residents, but often don’t know where to even begin looking. They also don’t always know just how critical the role of metadata is here, as only metadata can tell us where data comes from, where it resides in all the different systems, how it’s being used and by whom.
With Octopai’s automation tool, BI groups can auto-discover, access and retrieve metadata instantly in order to get complete data lineage, implement and track changes across systems (ETL, DB, DWH, Analysis and Reporting), determine and analyze impact of changes on reports, and identify privacy sensitive data immediately.
The role of machine learning in Octopai’s platform
Modeling & Indexing. Before we are able to upload an organization’s metadata to our centralized repository, we must first conduct modeling and indexing of the metadata. This process ensures all the metadata from each different system is in the same language (so to speak) so we can upload it to Octopai.
Determining dependencies. Once all the metadata has been indexed and uploaded to Octopai, it is important for us to determine existing correlations between all data elements. For example, we must determine dependencies that exist between data objects from within different systems and understand which systems are connected to which systems.
How do we do this?
Our machine learning algorithm is able to take many unrelated data elements and stitch them together to form more precise, complete data. It is able to identify the correlation between each of the data items that are extracted from a company’s systems and uploaded to Octopai’s platform, and augments that data with additional, external sources. Our algorithm identifies correlations that exist even between data items that don’t have the same name. For example, it can recognize different variations of the same word. Let’s take ‘Date of Birth’ for instance.
DOB, D.O.B., Date of Birth, dateofbirth, birthday, birthdate, Birth Date, dob, d.o.b…
There are so many variations used and these are just the ones in English. There might also be cases where a different language version is also used, or misspelled variations as well. Octopai takes all this into account and automatically identifies these correlations as being names for the same object or person etc. and groups them together. This is machine learning in metadata management, and significantly increases precision and accuracy in data governance.
It really is powerful
BI groups are feeling crushed by the endless amounts of data they must sift through regularly. They cannot easily find the data they need, understand it or leverage it to guide business decisions or comply with GDPR. Octopai offers vast and powerful capabilities – especially as organizations gear up for GDPR, and is empowering the BI group to tackle its challenges successfully and efficiently. In this big data universe, this giant robot machine monster (Octopai) is quite the formidable warrior to restore peace and calm to the BI group.
