With Dask, you can profile the compute graph and also visualize the profiler output. The Bokeh extension for JupyterLab lets you render those visualization within your notebook. For more details, see Dask Profiling.

ADS provides you a decorator that leverages the Dask profiler to help you visualize the CPU and memory utilization of supported operations.

How to Profile

Here is an example of how to profile the operation:

from ads.common.analyzer import resource_analyze
from ads.dataset.factory import DatasetFactory

def fetch_data():
    ds ="/home/datascience/ads-examples/oracle_data/orcl_attrition.csv", target="Attrition").set_positive_class('Yes')
    return ds

ds = fetch_data()

A graph is rendered on the notebook:

DS CPU and Memory Usage