October 11, 2022

HAROLD ROY ROJAS ESTRADA

Advertising campaigns play a crucial role in the success of businesses, and it is important to track their progress and measure their effectiveness. In this blog, we will take a deep dive into the final project selected of the Mojix Bootcamp, which was to create a dashboard report to provide insights and metrics on an advertising campaign.

Objective

Generate a Dashboard Report Campaign with insights and metrics.

Dataset

The dataset used for this project consisted on a collection of 70,993,469 instances (stored in a Parquet file) and 33 attributes. The dataset covered a month and a half of impressions from multiple campaigns (20 in total). Please note that the data is private property of Mojix Inc. and cannot be shared.

Tools

The following tools were used to create the Advertising Campaign Dashboard:

Procedure

  1. Data Ingestion: The first step was to mount the S3 bucket in Databricks, which was used to store the campaign data.
  2. EDA (Exploratory Data Analysis): SQL Queries and Pandas were used to analyze the data and uncover hidden patterns and insights.
  3. Data Viz: Databricks and Plotly was used to visualize the data and create interactive graphs.
  4. Dashboard: Databricks was also used to create a dashboard to display the insights and metrics in a user-friendly manner.
  5. Forecasting: The final step was to forecast the total impressions for the next 7 days using Prophet.

Databricks develops a web-based platform for working with Spark, that provides automated cluster management and IPython-style notebooks

Databricks develops a web-based platform for working with Spark, that provides automated cluster management and IPython-style notebooks

Procedure Diagram

Procedure Diagram

Code Examples

Results

The results of the Advertising Campaign Dashboard project were impressive. The dashboard provided a comprehensive analysis of the campaign, including: