As a data scientist your projects are invaluable to the business. They deliver revolutionary insights, paving the way for innovative new approaches within your organization. Yet there is often a barrier to making these insights repeatable and digestible throughout your organization.
Have you developed a project using Python, R or similar, but in order to continue to deliver great value to the business you are stuck re-running the same project? What you really want to do is move on to creating the next innovation.
Most business users use tools that query data using SQL such as Tableau, Qlik, Power BI etc. but your project is written in R, python, <insert your code of choice here>. In Kognitio, your code can be integrated into SQL objects that can be accessed directly from any ODBC/JDBC BI tool meaning that your analytics can be made available to the wider business without having to expose them to new complex interfaces.
Set up lightning fast reports or dashboards based on your analysis that can be accessed by the whole business. Kognitio’s external script interface can be used to execute your analysis and deliver results directly into Kognitio RAM. From here it can be access directly by any BI tool.
Provide analysis on-demand. Build a Kognitio external script object around your code: this is a SQL wrapper that controls the data input, execution and output format. Then, simply give BI and analytics tools access to the script in the same way as any other SQL object.
Take the pain out of on-demand analytics execution. Kognitio automatically manages data flow into multiple copies of your script and dynamically manages script parallelism to make best use of available CPU and RAM resources.
If you’re using Hadoop, you’ll know it uses various file formats like AVRO, Parquet etc. for different purposes and depending on your coding language of choice, you’ll need different ways to access these. There are many scenarios and each language/file-format combination may return a different structure. As a data scientist you want to spend your time analyzing data rather than doing arduous ETL wrangling to get your data into shape.
Kognitio simplifies this by providing your code with input data as CSV on the standard input stream and accepting output data as CSV on the standard output stream. This neatly offloads file format conversion and control of data flow to Kognitio. Prototyping is easy – simply use your favorite development environment. Just export test data from Kognitio to a CSV file and then pipe it into your code or connect directly using an ODBC/JDBC package.
What makes the worst bus stop in London?
Complex retail analytics on Hadoop
Learn more about Kognitio's data science consulting services
Big data sets – where to find and how to harvest them