Steps

Pipeline steps are connected together by their input and output and must form a unique path.

Therefore, you need to follow these principles:

• Pipeline must have at least one step without input as starting point. (e.g. a query step)
• Each input must be connected to one and only output. But same output can be used as input of multiple different steps.
• Depending on their type, some steps may not have inputs or outputs and may have specific rules to follow.

Data used and transfered during execution of a pipeline are not stored. Therefore you must use a plot script to produce some data visualization or a writer step to store data inside time-series and datapoints.

Each steps are connected together by named dataframes and can be from different types:

• Query step can be used to query time-series data and produce a Data Frame.
• Script step can be used to execute any Python function, e.g. a transformation or calling a third-party app.
• Model step defines how to train or interrogate a Machine Learning Model.
• Write step allows you to write back your results within Wizata.
• Plot step allows you to defines how you want to visualise your data.

One of the pipelines in the platform: