Harnessing KNIME Analytics

As part of my Cadetship at Spatial Vision, I designed and delivered an R&D project. After plenty of research, I chose to explore KNIME’s functionality, and how it can be applied within Spatial Vision. KNIME is a workflow based data transformation, analysis and deployment program.  Developed in Java using Eclipse as a free open source program, users are able to modify the base code of the program to suite their individual needs, allowing the creation of plugins to expand the functionality of KNIME. 

KNIME uses ‘Nodes’ to represent an input, process or output.  These nodes once connected to one another form a workflow.   Nodes are typically classified into 6 different categories based upon their purpose.

  • Data Access Nodes
  • Big Data Nodes
  • Transformation Nodes
  • Visualization Nodes
  • Deployment Nodes
  • Analysis Nodes

 

 Analysis and Data Mining is the other prime function of KNIME that really gives it an edge in terms of functionality.  Examples of these include advanced statistics, imagery analysis, web analytics, text mining, social media analysis and machine learning based prediction modelling. 

 

A unique aspect of KNIME’s nodes is that they can be packaged up into ‘Meta-nodes’. Meta-nodes are in essence a series of connected nodes that perform a set function but are wrapped up into their own individual node to assist reducing overwhelming complexity of workflows.

One of the key factors that makes KNIME so valuable is its ability to package up a workflow and distribute it to other users.  Inclusion of input data and any wrapped meta-nodes is also an option during exporting.  Workflows are small files and only become large if the input data is included.  If the recipient of the workflow has the same data then all they need to do is point the input node to their data location, hit run and the process would work identically. 

 

Overall, I have found KNIME to be a highly adaptable and valuable program.  It has the capacity to not only improve workflow methodologies but to also provide a mechanism for distribution of these workflows to clients, colleagues and affiliates. At this stage KNIME is not a perfect fit for spatial data sources so its use within Spatial Vision is limited however, the real value in KNIME is its machine learning capacity and ability to produce not only analytical results but drive innovation, integration and evolution of ideas to refine and improve upon existing and future methodologies.  KNIME is the type of program that is a real jack-of-all-trades and a master of none (to an extent).  This gap however is not exactly a weakness of the program itself but more of an opportunity for private developers to create their own nodes and tools to customise and improve the program to their own benefit.

For further information, please contact Spatial Vision at This email address is being protected from spambots. You need JavaScript enabled to view it.

 

Use the Buttons to Share this Blog:

Rate this item
(0 votes)

James Gordon

 

James was the Graduate Cadet for Spatial Vision in 2017. James worked over a range of projects during his first year with Spatial Vision which included the Jacaranda Atlas 9e; Electoral Re-Distribution Maps for the Queensland Electoral Commission and the Australian Electoral Commission; assisted with data migration services; helped Parks Victoria with mapping services and assisted with updating some of Spatial Visions primary products of VicMap Book, Outdoor Recreation Guide and Murray River Access Guide.

For any enquiries please contact us on 1300 36 67 96 or via contact form

 
Corporate Band
space Mapscape Space Check Site Space Gis Journey Space SV Maps Space Peri Urban

 

Login to post comments
Go to top