Theme
Description
Some of the experiment tasks in Python used language elements that participants considered hard to understand, even though they are commonly used in data pipelines written in Python / Pandas.
An example from task 2 is the addition of a boolean column:
data['publiclyFunded'] = data['traeger_art'].map(
lambda input: input == 'öffentlich'
)These advanced programming concepts are not immediately clear and force readers to either consult documentation or rely on comments if they have not previously encountered them.
Interestingly, users considered the blocks and pipe / blocktype concept in Jayvee to be intuitive, even though it is also an unusual programming concept (see also PL2.2 - Blocks structure improves understanding). This is potentially linked to the more direct mapping from the mental model of an ETL pipeline to blocks and pipes than to lambdas and map (HU2.1 - Mental model of an ETL pipeline aligns with Jayvee).
Representative Quotes
Some functions like lambda, list comprehension and implicit operations are not intuitive and require documentation and comments to understand.
- S12
Also, things like “map” and sqlite parameter like echo are not as crystal clear as the Jayvee oftype concept.
- S15
Lambdas are not really understandable to the layman.
- S26