Summary Python Machine Learning Tutorial: Fraud Detection System with Detailed Code Explanation

Machine Learning Code with Explanation (Fraud Detection Example)



Step 1️⃣: Importing Required Libraries

import numpy as np

➡ numpy is used for numerical operations (arrays, math, statistics).

import pandas as pd

➡ pandas is used to load and work with datasets in table format (rows & columns).

import matplotlib.pyplot as plt

➡ matplotlib helps in creating visual graphs and plots.

import seaborn as sns

➡ seaborn is used for advanced visualization; it makes graphs prettier and more readable.



Step 2️⃣: Load Dataset

df = pd.read_excel('/content/Fraud_Dataset_with_Issues.xlsx')

➡ Reads the Excel dataset and stores it in a DataFrame named df.



Step 3️⃣: Exploring the Dataset

df.head()

➡ Shows the first 5 rows of the dataset.

df.head(15)

➡ Shows the first 15 rows for more preview.

df.tail()

➡ Shows last 5 rows of the dataset.

df.tail(10)

➡ Shows last 10 rows of the dataset.

df.columns

➡ Lists all column names.

df.shape

, ➡ Shows number of rows and columns (format: rows, columns).

df.dtypes

➡ Displays the data types of each column (int, float, object/text).

df['CustomerLocation']

➡ Displays the “CustomerLocation” column values.

df['CustomerLocation'].unique()

➡ Shows unique values present in that column.

df.describe()

➡ Gives basic statistics summary like mean, min, max, std deviation for numeric columns.



Step 4️⃣: Handling Missing Values

df.isnull().sum()

➡ Counts missing (NULL) values in each column.

df = df.dropna()

➡ Removes all rows containing missing data.

df.shape

➡ Check new dataset size after removing missing values.

df.describe()

➡ New statistical summary after cleaning.



Step 5️⃣: Convert Text Labels to Numbers (Encoding)

df['CustomerLocation'] = df['CustomerLocation'].map({'Foreign': 1, 'Local': 0})

➡ Converts text categories to numbers because ML models require numeric input.

df.head()

➡ Check updated dataset.

df['CustomerLocation'].unique()

➡ Confirm conversion worked (should show only 0 and 1).

df['CustomerLocation'].value_counts()