Quick description of how to work with the data tag.
The data tag is used for initial reading purposes; it is the only required tag.
In every Nylon workflow, you must always provide a target column.
"data": {
"target": "ocean_proximity"
Here are the other optional grammar components you can include in your data section.

Data Reading Vocabulary

​Data Trimming​

trim: if you'd like to randomly get rid of a certain percentage of your dataset upon reading.

​Drop Columns​

drop: any columns you'd like to drop upon data reading.
"data": {
"target" "ocean_proximity",
"trim": 0.5,
"drop": ["name1", "name2"]

Custom Data Reading

Nylon supports custom data reading in case you're looking to pull your data from elsewhere. Regardless, the return format must be a Pandas DataFrame.
Nylon Specifications File
That's it for the data section. For any other pre-modeling data manipulation, head over to the preprocessing tab.