Data Cleaning Techniques

Data cleaning fixes errors and inconsistencies in a dataset, so the information going into your model is accurate and usable.

Common data cleaning tasks

• Removing duplicate rows.
• Fixing incorrect data types.
• Correcting inconsistent text formatting, such as mismatched capitalization.
• Removing irrelevant columns that do not help the model.

Fixing data types

Sometimes numbers get stored as text, which breaks calculations later.

import pandas as pd

df = pd.read_csv("data.csv")

df["Price"] = df["Price"].astype(float)

astype(float) converts the Price column from text to numbers, so you can use it in mathematical operations and feed it into a model.

Fixing inconsistent text

Text data often has mismatched casing, which can be treated as different categories by mistake.

df["City"] = df["City"].str.lower().str.strip()

str.lower() makes all text lowercase, and str.strip() removes extra spaces, so values like "Delhi" and " delhi " are treated as the same category.

Dropping unnecessary columns

df = df.drop(columns=["Unnamed: 0"])

This removes a column that adds no value to the analysis, like leftover index columns from a CSV file.

Clean data ensures a model learns from accurate patterns, not errors hidden in the dataset.