I was wondering if it would be possible to develop a connector for the open-source data analysis platform, KNIME. KNIME is a highly useful software for analyzing, transforming, and filtering data. It operates as a low-code platform, sharing similarities with Dynamo and Grasshopper. Those familiar with the latter tools could easily adapt to KNIME, making it accessible to a broader audience. Furthermore, there is already an extension available for importing IFC files in a tabular format, as discussed in this article.
I believe it would be beneficial to develop a connector that enables the integration of data from Speckle into KNIME. Numerous potential use cases come to mind, and I am interested in hearing your thoughts on this proposal.
I have used KNIME extensively in a previous life (processing brain scans, funnily enough) but not for a long time.
Are you actively using KNIME in your data workflows? Could you describe how Speckle connectivity would help as a specific connector? You can already make REST/GraphQL calls using OOTB nodes.
I mainly use it to do quality control checks and model analysis. I import the IFC in to the software and it’s very easy to manipulate the data to know if some attributes are empty or if the naming or format is correct. I can also very easily create graphics and reports and export them directly to Excel or Power BI.
You could also do most of this activities with Power Query or Power BI bringing the models form Speckle, but there’s the 100mb limit and that’s a very big handicap right now, also KNIME has a lot more flexibility and it’s open source.
A workflow that comes to mind is to import the IFC to KNIME, transform the data, add parameters and adapt the model to comply with certain quality levels; you could even make the model lighter and cleaner deleting unnecessary attributes, and then Send it to Speckle so it could be exchanged with other software ore people.
That’s just and example that comes to mind, but there are a lot of possibilities, and the workflows could be easily spread among the AEC professionals because of the similarities it has with Dynamo and Grasshopper. Right now we have the tools to do this activities with the proprietary formats (rvt, 3dm) but an open source and easy way to use no code with IFC is not that clear and I think this could be a great option.
Referring to the REST/GrapghQL, I’m a bit lost, I know they have something to do with API calls, but I’m not a developer so I’m not familiarized. If you could tell me about it referring to this topic it would be great!
Superficially only. As always, with the nodes and spaghetti paradigm, the data structures flowing down the pipelines are different shapes, as are the logical operations possible.
Interesting I came across Speckle and Knime in the same week. Was wanting to use knime to some how match object description to uniclass codes and ifc object type and then cross reference them to a costing database with older proprietary item descriptions. Here is sort of an example of doing a similar thing with powerBi and speckle in this video from the 4:30 mark. I am assuming with Knime I can automate the progress in a similar way. https://www.youtube.com/watch?v=kXVZqBlGv8Q
