aniruddha27

/* Sample data */
insert into emp (EMPID, NAME, JOB, SALARY)
values
(201, 'ANIRUDDHA', 'ANALYST', 2100),
(212, 'LAKSHAY', 'DATA ENGINEER', 2700),
(209, 'SIDDHARTH', 'DATA ENGINEER', 3000),
(232, 'ABHIRAJ', 'DATA SCIENTIST', 2500),
(205, 'RAM', 'ANALYST', 2500),
(222, 'PRANAV', 'MANAGER', 4500),
(202, 'SUNIL', 'MANAGER', 4800),

davidlenz

from sklearn.datasets import fetch_20newsgroups
import pandas as pd

def twenty_newsgroup_to_csv():
    newsgroups_train = fetch_20newsgroups(subset='train', remove=('headers', 'footers', 'quotes'))

    df = pd.DataFrame([newsgroups_train.data, newsgroups_train.target.tolist()]).T
    df.columns = ['text', 'target']

    targets = pd.DataFrame( newsgroups_train.target_names)

amberjrivera

If You've Never Used Sklearn's Pipeline Constructor...You're Doing It Wrong

How To Use sklearn Pipelines, FeatureUnions, and GridSearchCV With Your Own Transformers

By Emily Gill and Amber Rivera

What's a Pipeline and Why Use One?

The Pipeline constructor from sklearn allows you to chain transformers and estimators together into a sequence that functions as one cohesive unit. For example, if your model involves feature selection, standardization, and then regression, those three steps, each as it's own class, could be encapsulated together via Pipeline.

Benefits: readability, reusability and easier experimentation.