Welcome to Marketing common data modelling challenge!
At Improvado, we use marketing common data models (MCDM) to consolidate data from various ad platforms into a single source for our clients. MCDM is an imaginary model that maps fields from diverse datasets onto a single table or view, providing a comprehensive set of data for a particular report.
However, let's imagine a scenario where one of our MCDMs, responsible for reporting on paid ad performance, malfunctions. As a result, the client's dashboard displays inaccurate numbers for TikTok Ads, which is unacceptable since TikTok is a crucial platform for our client.
Adding to the issue, the analyst responsible for this particular MCDM is currently on vacation. As a result, we are left with only raw data from the ad systems, the MCDM table structure for the client's report, and a broken dashboard.
In this situation, we need you to:
- Begin a new project in dbt Cloud, utilizing Google Big Query as the DWH.
- Use the raw data from the ad platforms and the MCDM table structure for the ads_basic_performance report to develop a corrected MCDM model:
- The model must include data from all the provided ad platforms.
- The model's columns must match the structure of the provided MCDM table.
- The model should be simple to update with any new ad platforms that may be introduced.
- Recreate the broken client's dashboard using Google Looker Studio.
To submit the task, please send an email to vladimir.kobzev@improvado.io with the following completed steps:
- A link to your dbt Cloud repository that contains the completed MCDM for the ads_basic_performance report.
- A link to the fixed dashboard.
- A brief set of instructions for adding data from new ad platforms into your MCDM.
To complete this task, you will need the following tools:
- dbt Cloud
- Google Big Query
- Google Looker Studio
To help you get started, here are some resources on how to use the necessary tools:
- dbt Courses:
- dbt Fundamentals. Relevant chapters include:
- Setting up dbt Cloud (17 minutes)
- Models and Sources (40 minutes)
- dbt Cloud and BigQuery for Admins (35 minutes)
- dbt Fundamentals. Relevant chapters include:
- How to Use Google BigQuery for FREE
- How to create dashboard Google Looker Studio with Google Big Query
- Learn more about dbt in the docs
- Short overview for Improvado MCDM
This is the foundational repository for your project. Clone it and start your dbt Cloud from it.
The repository includes raw data from various ad platforms, as well as the MCDM structure for the ads_basic_performance report, which are provided as seeds:
- src_ads_bing_all_data
- src_ads_creative_facebook_all_data
- src_ads_tiktok_ads_all_data
- src_promoted_tweets_twitter_all_data
- mcdm_paid_ads_basic_performance_structure
To build the seeds, run dbt seed
in the dbt Cloud console. Once the seeds have been built, you can access the data using ref()
. For example, you can use select * from {{ ref('src_ads_bing_all_data')}}
to access data from the src_ads_bing_all_data
seed.