What Is The Biggest Difference Between Alteryx and KNIME?

Although both Alteryx and KNIME can be found in the 2019 Gartner Magic Quadrant for Data Science and Machine Learning Platforms, and they do have in common the paradigm of a visual workflow, they do have differences. Alteryx has a well-earned reputation for letting the user easily connect to different kinds of data sources and bring them into a workflow to create datasets for use in analysis. They are very good at this, and I've heard engineers speak well about that ability, preferring to use Alteryx for their ETL than to learn ways of doing the same work in SQL Server. Alteryx bought yhat, a data science platform, and have been gradually adding capabilities for predictive analytics into the Alteryx workflow. My own experience trying a data science predictive project in Alteryx showed me that the company still has a long way to go. The same predictive project was far slower with Alteryx on my laptop than using R code (in RStudio) on my laptop, and there were not many tools in Alteryx that could do what I could do with R coding. At least Alteryx does allow both R and Python script blocks of your own creation to be inserted as steps in the workflow. But, in my opinion, Alteryx is definitely NOT a suitable data science platform, as of this writing.

KNIME, on the other hand, is a more complete data science platform. I have used it just enough to get a feel for what it can do. It also allows your own code blocks as part of the workflow. I can't recall if it is R, Python or both. Also, like Alteryx, KNIME has developed connectors for many kinds of data sources. I have not tried that same R project in KNIME that I did in Alteryx, so I can't really compare its ability as a data science platform to the Alteryx. I would expect that Alteryx has a free trial version, and I know that KNIME is free software (they make money off of things like KNIME Server). If I were to recommend one or the other, I would recommend KNIME. But, if your company has standardized on Alteryx, you'll have a hard time convincing them to let you use KNIME. If money is not an issue, I would look for a data science platform that does the job better than either of these. I would consider RapidMiner, Dataiku and DataRobot. You can see the movement of all the above-mentioned tools from Gartner's placement of them in the 2018 and 2019 magic quadrants here: towardsdatascience.com