Named Entity Recognition — Hands-on

NER with spaCy (Industry-Standard, Pre-trained)

import spacy

nlp = spacy.load("en_core_web_sm")
text = "Riya joined Google in Bengaluru on 5 August 2026 for a package of ₹18,00,000."
doc = nlp(text)

for ent in doc.ents:
    print(ent.text, "->", ent.label_)
# Riya            -> PERSON
# Google          -> ORG
# Bengaluru       -> GPE
# 5 August 2026   -> DATE
# ₹18,00,000      -> MONEY

Visualizing Entities

from spacy import displacy

displacy.render(doc, style="ent")   # renders inline HTML highlighting each entity

NER with NLTK (Chunking-based)

import nltk
from nltk import pos_tag, word_tokenize, ne_chunk

text = "Riya joined Google in Bengaluru."
tree = ne_chunk(pos_tag(word_tokenize(text)))
print(tree)
# (S
#   (PERSON Riya/NNP)
#   joined/VBD
#   (ORGANIZATION Google/NNP)
#   in/IN
#   (GPE Bengaluru/NNP)
#   ./.)

Extracting Only PERSON and ORG Entities

import spacy

nlp = spacy.load("en_core_web_sm")
text = "Sundar Pichai leads Google, while Satya Nadella leads Microsoft."
doc = nlp(text)

people = [ent.text for ent in doc.ents if ent.label_ == "PERSON"]
orgs = [ent.text for ent in doc.ents if ent.label_ == "ORG"]
print("People:", people)   # ['Sundar Pichai', 'Satya Nadella']
print("Orgs:  ", orgs)     # ['Google', 'Microsoft']

Building a Simple Gazetteer-based (Rule) NER for a Custom Domain

Pre-trained NER models won't recognize domain-specific entities (e.g. course names, internal product codes). spaCy allows adding custom entity rules:

import spacy
from spacy.pipeline import EntityRuler

nlp = spacy.load("en_core_web_sm")
ruler = nlp.add_pipe("entity_ruler", before="ner")

patterns = [
    {"label": "COURSE", "pattern": "Natural Language Processing"},
    {"label": "COURSE", "pattern": "Data Structures"},
]
ruler.add_patterns(patterns)

doc = nlp("This semester I am studying Natural Language Processing and Data Structures.")
for ent in doc.ents:
    print(ent.text, ent.label_)
# Natural Language Processing -> COURSE
# Data Structures             -> COURSE

Real-World NER Applications

NER output (structured entity tags) directly feeds into Information Extraction systems, which we cover in Unit 3.