Analyzing Digital Marketing Campaigns with Data Science
Posted: Sun Dec 22, 2024 6:00 am
Harve's Data Science course was my first contact with the data science area and the project was also my first. The initial idea of the final Data Science project was to create a Marketing Mix Modeling where the MMM's function is to understand and evaluate the performance of the campaigns and understand how much more effective each one was during the period of time in which it was carried out.
So the objective of the project is to identify and understand which advertising channel is bringing the greatest return on investment (ROI), thus being able to more assertively optimize the campaign budget and, at the same time, create an increasingly appropriate and effective strategy
Project execution:
So the chosen database is a database found on GitHub, which french email address list information on campaigns from a company in Hong Kong, China, such as demand, sales, advertising expense data, investment in SMS, newspaper, radio, TV, internet and gross audience points (GRP) for each channel invested in, thus making it possible to calculate the effectiveness of the campaigns.
Follow the steps of the project carried out at Cola
1st Importing the libraries and extracting the .xlsx file from the drive where it was located
Then the data processing and transformation steps begin.
Preview first 5 lines
Number of rows and columns
Column description
Column information
Find out if there are missing values in the columns
Renaming some columns
Creating a list from another list, removing the columns that I considered unnecessary
In this way, the data analyzed were only those shown in the image below.
From this data it was possible to obtain answers to some questions such as:
Which channel had the highest GRP per year?
Which channel had the highest cost per year?
Which channel generated the most demand?
This table was developed so that we can understand that during all the years from 2010 to 2017 the investments were linear, they remain similar, there was no drop in investment, however, even maintaining almost the same investment value, it was possible to notice a large drop in sales over the years, the colored lines are the investments and the bars are the sales during each year.
So the objective of the project is to identify and understand which advertising channel is bringing the greatest return on investment (ROI), thus being able to more assertively optimize the campaign budget and, at the same time, create an increasingly appropriate and effective strategy
Project execution:
So the chosen database is a database found on GitHub, which french email address list information on campaigns from a company in Hong Kong, China, such as demand, sales, advertising expense data, investment in SMS, newspaper, radio, TV, internet and gross audience points (GRP) for each channel invested in, thus making it possible to calculate the effectiveness of the campaigns.
Follow the steps of the project carried out at Cola
1st Importing the libraries and extracting the .xlsx file from the drive where it was located
Then the data processing and transformation steps begin.
Preview first 5 lines
Number of rows and columns
Column description
Column information
Find out if there are missing values in the columns
Renaming some columns
Creating a list from another list, removing the columns that I considered unnecessary
In this way, the data analyzed were only those shown in the image below.
From this data it was possible to obtain answers to some questions such as:
Which channel had the highest GRP per year?
Which channel had the highest cost per year?
Which channel generated the most demand?
This table was developed so that we can understand that during all the years from 2010 to 2017 the investments were linear, they remain similar, there was no drop in investment, however, even maintaining almost the same investment value, it was possible to notice a large drop in sales over the years, the colored lines are the investments and the bars are the sales during each year.