Credit Card Fraud
This dataset consists of credit card transactions in the western United States. It includes information about each transaction including customer details, the merchant and category of purchase, and whether or not the transaction was a fraud.
I will use this dataset to build a model to predict whether a credit card transaction was fraudulent or not.
Below is the dataset:
import pandas as pd
df = pd.read_csv('credit_card_fraud.csv')
dfData Dictionary
| transdatetrans_time | Transaction DateTime |
|---|---|
| merchant | Merchant Name |
| category | Category of Merchant |
| amt | Amount of Transaction |
| city | City of Credit Card Holder |
| state | State of Credit Card Holder |
| lat | Latitude Location of Purchase |
| long | Longitude Location of Purchase |
| city_pop | Credit Card Holder's City Population |
| job | Job of Credit Card Holder |
| dob | Date of Birth of Credit Card Holder |
| trans_num | Transaction Number |
| merch_lat | Latitude Location of Merchant |
| merch_long | Longitude Location of Merchant |
| is_fraud | Whether Transaction is Fraud (1) or Not (0) |
Source of dataset. The data was partially cleaned and adapted by DataCamp.
Creating a DataFrame with only fraudulent transactions
f = df[df['is_fraud'] == 1]
fprint(df.shape)
print(f.shape)Introduction
As the outcome is a classification problem with only two options ( fraud or not fraud) a classification models will be built and tested. In this instance I will use a Decision Tree Classifier.
Before building the predictive model, I will explore the data, looking for patterns, correlations and trends in the data. Visualizations will be used to better understand and visualize the makeup of the data.
The data in the dataset will also need to be prepared for the model. All categorical fields will need to be changed to numeric, some models will require scaling and any missing values will need to be dealt with.
After the exploration and preparation the model will be built and optimized using training data. Then, the model will be tested using never seen data.
After it is created, a visualization of the decision tree will be shown.
Importing Basic Libraries
import pandas as pd
import numpy as npData Exploration
In this section I will explore the data, looking at basic statistics, the size, shape and makeup of the dataset, distributions, and correlations. Some of these will be text-based explorations while others will be visuals.
The purposes of this section is to get a better feel and understanding of the data. To see if there are any immediately apparent patterns or trends, and to get a better sense if our final results align with our initial impressions.
Distributions
I will start by exploring and visualizing the distributions of the dataset
#Libraries
from matplotlib import pyplot as plt
from matplotlib.ticker import PercentFormatter
import foliumDistribution of Fraud