Business Area
Marketing analytics and campaign performance
Data Analytics Case Study
Marketing Campaign Performance Analysis
Evaluating campaign performance across channels, customer segments, companies and time using more than 200,000 marketing records
Python • Pandas • NumPy • Matplotlib • Seaborn • Tableau • Exploratory Data Analysis • Business Reporting
This project analyses marketing campaign performance across multiple business dimensions, including campaign type, marketing channel, customer segment, company performance, acquisition cost, engagement, impressions, clicks, conversion rate and return on investment.
Project Overview
Dataset Size
More than 200,000 campaign records
Primary Goal
Evaluate campaign performance and identify meaningful optimisation opportunities
Main Tools
Python, Pandas, Matplotlib, Seaborn and Tableau
Business Problem
Marketing teams often manage campaigns across several channels, customer segments and campaign formats. Decision-makers need a clear way to compare performance and understand whether differences in ROI, conversion rate, engagement and acquisition cost are large enough to support strategic action.
This project evaluates the dataset from multiple perspectives to determine whether specific campaign types, channels, customer segments or companies consistently outperform the others.
Key Business Question
Which campaign characteristics are associated with stronger marketing performance, and what recommendations can be supported by the available evidence?
Dataset
The dataset contains more than 200,000 marketing campaign records. Each row represents a campaign observation with performance metrics, categorical descriptors and a reporting date.
Campaign Dimensions
Company, campaign type, marketing channel and customer segment
Performance Metrics
ROI, conversion rate, clicks, impressions, engagement score and acquisition cost
Time Variable
Campaign date used for monthly performance analysis
Analysis Type
Descriptive, exploratory and correlation analysis
Analytical Process
The project followed a structured workflow from defining the business question to communicating insights through dashboards and executive reporting.
Ask
Defined the business problem and analytical objectives.
Prepare
Loaded the dataset and reviewed its structure and fields.
Process
Validated data quality, duplicates, dates and missing values.
Analyze
Compared KPIs across campaigns, channels and segments.
Visualize
Created Python charts, a heatmap and Tableau dashboards.
Report
Developed findings, recommendations and an executive summary.
Behind the Analysis
This section connects the Python workflow to its analytical purpose, output and business value. Open each step to explore the process.
STEP 01 Import Libraries Pandas, NumPy and Matplotlib were prepared for data handling, analysis and visualisation.
Importing the Required Python Libraries
The Python environment was prepared with the libraries required to load, inspect, aggregate and visualise the data.
import pandas as pd
import numpy as np
import matplotlib.pyplot as pltPurpose
Prepare the tools needed for the analytical workflow.
Why It Matters
Pandas supports data preparation and aggregation, while Matplotlib supports chart creation.
Business Value
These tools provide the foundation for converting raw campaign records into decision-ready insights.
STEP 02 Load the Data The marketing campaign dataset was loaded into a Pandas DataFrame and its size was confirmed.
Loading the Marketing Campaign Dataset
The source dataset was imported into Python to begin the inspection and analysis process.
STEP 03 Explore the Dataset Sample records, data types, structure and missing values were inspected before preparation.
Understanding the Dataset Structure
The dataset was inspected to understand its dimensions, column types, sample values and potential quality issues.
df.shape
df.head()
df.info()
df.isnull().sum()Dataset Preview
Dataset Information
Missing Values
STEP 04 Validate and Prepare Duplicate records were assessed and the Date column was converted to a datetime format.
Data Validation and Preparation
The dataset was checked for fully duplicated rows and the reporting date was converted into a format suitable for time-based analysis.
Assess Duplicate Records
Fully duplicated rows were counted to prevent totals and averages from being overstated.
Confirm Duplicate Finding
No fully duplicated records were identified, so all campaign rows were retained.
Convert the Date Column
The Date field was converted to datetime format for monthly trend analysis.
Validate the Data Type
The updated data type was checked to confirm that the conversion was successful.
STEP 05 Exploratory Data Analysis Campaign performance was compared across campaign types, channels, customer segments, companies and time.
Comparing Marketing Performance
Grouped summaries and visualisations were used to compare ROI, conversion rate, engagement, acquisition cost and monthly performance.
Average ROI by Campaign Type
Observation
Average ROI remained close to 5 across all campaign types, with only very small differences.
Business Insight
No campaign type demonstrated a meaningful financial advantage within this dataset.
Average Conversion Rate by Marketing Channel
Observation
Average conversion rates were highly consistent across the available marketing channels.
Business Insight
Channel selection should be evaluated alongside cost, reach and engagement rather than conversion rate alone.
Average Engagement Score by Customer Segment
Observation
Engagement scores were distributed evenly across customer segments.
Business Insight
No segment displayed substantially stronger engagement, so segmentation decisions require additional behavioural information.
Average Acquisition Cost by Company
Observation
Average acquisition costs were relatively similar across all five companies.
Business Insight
The observed differences were too small to support major budget-allocation decisions by company.
Monthly Average ROI
Observation
Monthly average ROI remained stable throughout the reporting period.
Business Insight
No significant seasonal trend or sustained monthly performance shift was identified.
Correlation Analysis
Observation
Most correlation coefficients were close to zero, indicating very weak linear relationships.
Business Insight
Marketing performance should be assessed through multiple KPIs rather than relying on one variable as a predictor of another.
STEP 06 Build Tableau Dashboards Executive and trend dashboards were created to support business monitoring and decision-making.
Communicating Results Through Tableau
Two dashboards were developed to present key performance indicators, comparisons and time-based trends in a concise, business-friendly format.
Executive Dashboard
Marketing Trends Dashboard
Visualizations
The visual analysis compares campaign performance across key marketing dimensions and tests whether meaningful differences exist.
ROI by Campaign Type
Compares the average return on investment generated by each campaign category.
Conversion Rate by Channel
Evaluates conversion performance across different marketing channels.
Engagement by Segment
Compares average customer engagement across audience segments.
Acquisition Cost by Company
Compares customer acquisition costs across the companies in the dataset.
Monthly ROI Trend
Examines how average return on investment changed over time.
Correlation Heatmap
Measures the direction and strength of linear relationships among numerical marketing variables.
Tableau Dashboards
The dashboards translate the Python analysis into an accessible reporting format for business users and decision-makers.
Executive Marketing Dashboard
Marketing Trends Dashboard
Additional Exploratory Analyses
Additional analyses were completed in the full notebook to validate the findings from multiple perspectives.
Average Clicks by Marketing Channel
Click volumes were compared across channels to evaluate audience response and traffic generation.
Average Impressions by Campaign Type
Campaign reach was compared across campaign categories.
ROI and Engagement Distributions
Distribution analysis was used to inspect spread, central tendency and possible extreme values.
Top Companies by Average ROI
Company-level ROI rankings were reviewed, although the observed differences remained small.
Key Findings
The analysis found that most campaign performance measures were highly balanced across categories.
Campaign ROI was highly consistent.
Average ROI remained close to 5 across all campaign types, with no category showing a significant financial advantage.
Marketing channels performed similarly.
Average conversion rates and click volumes were consistent, with no individual channel significantly outperforming others.
Customer engagement was evenly distributed.
No customer segment demonstrated substantially higher engagement levels.
Company performance differences were small.
Acquisition cost and ROI were relatively similar across all five companies.
Monthly performance remained stable.
No significant seasonal trend or sustained fluctuation in average ROI was identified.
Linear relationships were very weak.
Most correlation coefficients were close to zero, suggesting that the numerical variables operated largely independently.
Business Recommendations
The findings support a balanced and multi-metric approach to marketing performance management.
Use Multiple KPIs
Evaluate campaigns through ROI, conversion rate, engagement, acquisition cost, clicks and impressions instead of relying on a single performance measure.
Continue Performance Monitoring
Track campaign performance over time to identify future changes that may not be visible in the current balanced data.
Add Richer Business Variables
Expand future analysis with variables such as revenue, profit, campaign duration, budget, geographic market and customer lifetime value.
Use Interactive Dashboards
Maintain Tableau dashboards to support ongoing reporting, filtering and faster management decision-making.
Executive Summary
This project analysed more than 200,000 marketing campaign records using Python to evaluate performance across campaign type, marketing channel, customer segment, company, acquisition cost, engagement, impressions, clicks, conversion rate and return on investment.
The analysis found that the dataset was highly balanced. Most campaign types, channels, segments and companies produced similar average results, while monthly ROI remained stable and numerical variables showed very weak correlations.
Although the dataset did not reveal strong performance differences, the project successfully demonstrates a complete exploratory data analysis workflow, including data loading, validation, preparation, aggregation, visualisation, statistical analysis, business interpretation, dashboard development and executive reporting.
Skills Demonstrated
Python
Pandas
NumPy
Matplotlib
Seaborn
Tableau
Data Validation
Exploratory Data Analysis
Data Aggregation
Correlation Analysis
Marketing Analytics
Business Reporting
Data Visualization
Business Recommendations
Project Resources
Explore the complete project through the Kaggle notebook and interactive Tableau dashboards. These resources include the full Python workflow, data preparation, exploratory data analysis, statistical insights, executive dashboards, and business reporting.