Data-projects-with-R-and-GitHub

Social Media Advertising and Campaign Performance

Social media advertising is an important part of modern marketing and public relations. This project explores how social media advertising campaigns differ across platforms, customer segments, and campaign goals.

The focus of this project lies on cleaning, restructuring, and visualizing advertising campaign data. I am especially interested in how different advertising strategies lead to audience interaction and financial performance.

Dataset

The original dataset can be found on Kaggle, which is an online platform where users can publish, share, and download datasets for data analysis projects.

https://www.kaggle.com/datasets/jsonk11/social-media-advertising-dataset/data

However, since Kaggle requires user registration to download the dataset, a local copy of the CSV file can be found within ‘Minseo’ project folder.

Summary of the Dataset

Campaign_ID Target_Audience Campaign_Goal Duration Channel_Used Conversion_Rate Acquisition_Cost ROI Location Language Clicks Impressions Engagement_Score Customer_Segment Date Company
529013 Men 35-44 Product Launch 15 Days Instagram 0.15 $500.00 5.7900000 Las Vegas Spanish 500 3000 7 Health 2/25/2022 Aura Align
275352 Women 45-60 Market Expansion 15 Days Facebook 0.01 $500.00 7.2100000 Los Angeles French 500 3000 5 Home 5/12/2022 Hearth Harmony
692322 Men 45-60 Product Launch 15 Days Instagram 0.08 $500.00 0.4300000 Austin Spanish 500 3000 9 Technology 6/19/2022 Cyber Circuit
675757 Men 25-34 Increase Sales 15 Days Pinterest 0.03 $500.00 0.9098236 Miami Spanish 293 1937 1 Health 9/8/2022 Well Wish
535900 Men 45-60 Market Expansion 15 Days Pinterest 0.13 $500.00 1.4228282 Austin French 293 1937 1 Home 8/24/2022 Hearth Harmony
323031 Women 35-44 Product Launch 15 Days Facebook 0.02 $500.00 6.9000000 Austin Spanish 500 3001 10 Technology 1/15/2022 Cyber Circuit

The dataset contains information about social media advertising campaigns, including audience targeting, campaign goals, platforms, engagement scores, clicks, impressions, acquisition costs, and ROI.

The data is provided as a CSV file and includes the following variables:

Variables

Variable Description
Campaign_ID Unique identifier of the advertising campaign
Target_Audience Demographic group targeted by the campaign
Campaign_Goal Objective of the campaign
Duration Campaign duration written as text values
Channel_Used Social media platform used for the campaign
Conversion_Rate Percentage of successful conversions
Acquisition_Cost Customer acquisition cost
ROI Return on investment
Location Campaign location
Language Language used for the campaign
Clicks Number of advertisement clicks
Impressions Number of advertisement impressions
Engagement_Score Standardized audience engagement score
Customer_Segment Targeted customer category
Date Campaign date
Company Company running the campaign

Tasks for Data Clean Up

Before the data can be analyzed, several variables need to be cleaned or restructured.

These steps are important because several variables are not immediately ready for numerical analysis. For example, campaign duration is stored as text, acquisition cost contains currency symbols, and the target audience combines multiple demographic categories in one column.

Tasks for Data Manipulation

1. Which campaign goals bring the highest engagement scores?

One important question is which campaign goals are most effective at generating audience interaction on social media.

To answer this question, the data will be grouped by Campaign_Goal. For each campaign goal, create a nicely formatted summary table showing:

The standard deviation should help show whether engagement scores are consistent within each campaign goal or vary strongly between campaigns. Also, this overall table should help identify whether certain campaign goals are more successful at attracting audience involvement.

2. Which age group has the highest click-through rate?

Another important question is which age groups are most likely to interact with advertisements after seeing them.

Instead of only comparing raw click counts, I will calculate a click-through rate.

Click-through rate will be calculated as Clicks / Impressions.

To make the analysis more precise, each age range should be decomposed into individual ages. For example, the campaign targeting 25-34 should be counted for every age from 25 to 34. After restructuring the data caculate the average click-through rate for each individual age! To answer this question, the Target_Audience column should first be split into separate demographic variables. Then the average click-through rate will be calculated for each age group.

This makes the comparison fairer than using only raw click counts, because campaigns with more impressions naturally tend to have more clicks.

Tip: Check whether the Target_Audience column has been properly split into separate gender and age group variables.

Additional task of data visualization

As an additional challenge, a small data visualization could also be created here:

Create a age-separated click-through rate plot using faceting by gender. The plot should show click-through rate by exact age separately for male and female target audiences.

Tip: Campaigns labeled as targeting “all” should be counted for both men and women.

Tasks for Data Visualization

Create a line graph showing how average ROI changes with campaign duration across different social media platforms.

The data should be grouped by Duration and Channel_Used.

The final graph should show:

This visualization should help explore whether longer advertising campaigns are associated with higher ROI, and whether this pattern differs across social media platforms.

Tip: Check whether the Duration column has been converted into numeric day values before creating the graph.