Alteryx is a powerful software capable of performing different tasks. One can create a full Data Science pipeline from data collection, data engineering, modelling using available models, calculate the metrics of these models and compare these models to each other. In addition, if you want to try new algorithms that are not available in Alteryx yet but the Python already has them, you should try the Python tool. In other words, you can integrate your Python script into your Alteryx workflow, and we will consider how to do so in this post.
A developer can find the Python tool by typing python in the search field. By pressing on the Python tool, you obtain the following interface:
You can see the Python tool on the Canvas, and on the Configuration window, you can find Jupyter Notebook. The developer can run cells within Jypyter Notebook the same way as Jupyter notebook is not in Alteryx using button run and others. To install Python libraries, one needs to uncomment the following line: #Package.installPackages([‘pandas’,’numpy’]) and add the desired Python library. For example, to add the sklearn library, it is necessary to write: Package.installPackages([‘pandas’,’numpy’, ‘scikit-learn]’).
To read a dataset from the incoming connection, one needs to write:
df = Alteryx.read(‘#1’)
where #1 is the number of incoming connections. To write data to one of five output anchors, a developer needs to write:
where df is the dataset to output and 2 is the anchor number.
Alteryx gives you a very flexible opportunity to add Python script into the Alteryx workflow compared to competitors. You may not need the Python tool when you work with data on a daily basis, but it is a good tool to have in your arsenal.