Every now and then I have a chance to talk to someone about NNCompass. Nine times out of 10 it goes exceedingly well! That tenth time though, well, let’s just say I need to work on either my mastery of the vernacular, or… I need to do a better job of conveying its features up front. Usually at the end of the meeting, where I’ve just poured my heart and soul into describing NNCompass, which I’ll admit can be frantic and all over the map sometimes, they’ll hit me with this question… “so what exactly does NNCompass do”? At which point, I realize, I have gone way too far into the weeds… So, here’s my attempt at a mea culpa blog post, where I try and abstract a quick description of what NNCompass is, and what its core features are.
If you could describe NNCompass quickly, how would you do that?
NNCompass changes the paradigm of data management, making the techniques to construct Smart Data uncomplicated and accessible to everyone. You just don’t need teams of expensive data engineers or data scientists, and rows of server racks anymore to gain insights and use tools like machine learning. And not just structured data like CSV or XLS – NNCompass allows you to take advantage of all your data, including insights hidden away in unstructured text.
Explain that a little more?
NNCompass represents a new breed of data management software that uses human intelligence in concert with machine learning to transform “raw” data into effective and actionable ”Smart Data”. Everyone should be able to create Smart Data; you shouldn't need to be a software engineer or data scientist to take advantage of algorithms that will put your data to work for you. Some of the most valuable insights are locked away in unstructured text, like MS Word, PDF, emails and social media comments. NNCompass allows you to get at all of the data you would typically find in a business environment.
What are the core features of NNCompass?
- Data Ingest: NNCompass allows the user to upload and connect to data either on-premise or in the cloud. All of this is done in-memory, so you don’t need expensive data engineering teams, huge technical footprints or rows of server racks to do this. When a user connects to data, we run algorithms against the data to provide insights into everything from inferring the types of data, to insights into where your data is similar and could be integrated, and streamlining into what we call a Unified Data Format (UDF).
- Data Wrangling: Making sure your data is cleansed, normalized and wrangled are important steps in the data management workflow. NNCompass allows you to apply a wide array of transforms, all done in-memory, so you can see the results of those wrangling transforms immediately.
- Data Integration: Creating groups of data in NNCompass centered around a concept couldn’t be easier. Once you’ve grouped data together, you can select fields to integrate your data on, effectively linking the data together and creating a unified picture.
- Data Streaming: There are dozens of input/output connectors available in NNCompass to take advantage of large or streaming data sets. Once you’ve uploaded example files and performed your wrangling and transform operations, you can set a stream up that will apply all of those transforms to the thousands or millions of files in your data stream.
- Unstructured Document Handling: The truth is most business are dealing with large volumes of unstructured data, whether that’s word documents, emails, PDF’s,… a lot of tools on the market do a great job of managing highly structured data, but what about the other 80% of information that you could be taking advantage of? We needed NNCompass to help users incorporate all of their data, and through the use of point-and-click interfaces, extract knowledge and apply structure to unstructured text… then easily integrate that with other data to give you a true 360 degree view of your information.
- Smart Data Workbench: We talk about leveraging human intelligence on top of, or in concert with machine learning. Making the process of creating smart data, data that is highly valuable, accessible and attainable, and NNCompass was designed along this principle. That you can take someone, who isn’t technical, but has in depth knowledge of your data and questions that need to be answered, and allow that subject matter expert to employ algorithms, machine learning, deep learning in a simplified way to create value, not just from CSV and Excel files, but from all the data you would typically find in a business environment.
What about NNCompass Enterprise?
NNCompass Enterprise includes all the core features described above, can deployed at a customer site, allows users to collaborate on data by sharing data sources, data groups and data wrangling operations (transforms).
What about NNCompass Government?
Everything in NNCompass Enterprise, and NNCompass Government implements custom processing, parsing and streaming code to handle some of the most difficult government message types (matter of fact, 465 different government message types), data formats and security tagging environments. These capabilities allow your team of analysts to focus on extracting value out of the data, as opposed to working endlessly to ingest, parse and ensure proper security tags on your data. Our custom security tagging code is used to parse and verify any text that we identify in a document as a candidate marking. We handle dozens of government message types, automatically applying structure making it easy to wrangle, transform, and importantly, integrate all of your data sources together.
As I finish writing this I realize I just wrote close to 1000 words… so maybe not the best attempt of “abstraction”, but stick with me! I’m getting better at this whole message conveyance conundrum!