Knime DB

Hi all

I am new in KNIME and I need some ideas 

I have an app for data analytics. Basicalley, in this app some data is analysed and a dashboard is already there. I want to use this data for prediction purpose. And I want to allow to users to ask for the prediction results at anytime. We will manage this via a user interface.

To achieve this, I think about integrating KNIME into the app so my idea is to

  1. Integrate KNIME
  2. Modeling for prediction
  3. Via the user interface, allow the user can jit on a button to run the model and receive some prediction

My question is:

Is it possible to integrate knime ?

Or how to integrate our  DB->KNIME.  Perhaps there is a sql interface available in KNIME. But is another staging area in some server needed for storing the results (e.g. datasets) in right format? Where for example are all the internal KNIME data, models etc?

And server KNIME? Is there this kind of option available (Rapid Miner has that option). Because if yes, it potentially will give more freedom to also for model deployment. E.g for user to run what-if analysis? And if server would be possible, is own GUI for that possible?


Thank you 



You can always create a KNIME model through a KNIME Workflow. Then apply the model (PMML maybe?) to some new data. You can score the model via PMML libraries or export the whole workflow as API. You will need the KNIME Server for the latter.

Please check the following use case and the following blog post

Does this answer your question?


Thanks !

What about Knime server ?

I mean when I import data and process my model, when I will store the results, the models, the data ? I mean will I need a sever ? or these can be stroed in some place in KNIME ?

For example lite version  has shared db and web portal access to model streams. Does it mean some sort of model deploment?

The KNIME Server provides shared space for workflows, data, metanodes and models.

The WebPortal is one of the many deployment possibilities when a web interface is required.

Another deployment possibility could be exporting the workflow as an API service (server required).

Another deployment possibility is just to run the workflow scheduled or on demand, remotely or locally.

For remote  and scheduled execution the server is required.