Introduction to Pipelines

Traditionally, this step would be performed in Python or SQL code and it is where we come to model & cleanse a dataset to optimise its quality for analysis.

1. Begin by selecting ‘create a new pipeline’. This will navigate you to the pipeline creation page where you will start by selecting ‘configure’ to choose the base dataset that you wish to transform. We can quickly preview the dataset to make sure that it looks as we expect it to look, before taking further steps.
2. Once selected, the dataset will load and a green tick in the top right corner of the box will appear to confirm its readiness.
3. Click, drag and drop a transformation step (the box on the top left of the page) to just below your base dataset so that it snaps into place.
4. Configure the transformation step that you wish to apply (note, there are multiple transformation steps that can be made at this stage, see above for those transformations and the cases in which they should be used). Depending on the type of transformation you have selected, the fields to be completed in the box will vary. Save this step once completed.
5. The toggle will now appear in the top right of the transformation step box. This represents whether the transformed dataset will be saved to your datasets tab, and will be toggled off by default. Note - it is best to only toggle on if you intend to perform an analysis on that particular dataset as this will then save the final dataset for visualisation.

Steps 3 & 4 can also be performed with an AI model in lieu of a transformation. Simply repeat the above steps, select ‘configure’ and choose the type of AI model that you wish to apply (Forecast, Group, Missing Values or Reducing Complexity). Be sure to save the AI model step so that the dataset is ready for visualisation in the next stage.