Building a data warehouse (DWH), training a machine learning (ML) model, and creating a Power BI dashboard A-DWH Project Requirements:
- Identify Bus matrix for your business processes against common dimensions
- State the dimensional modeling process for each business process
- Using any diagramming tool you like, construct a logical data model for this case study. Output should be an image or PDF. State why did you choose this particular data model design? What does the data represent? (Details about each model component is necessary).
- Translate the logical data model to a physical data model which includes the following: tables and columns (name, data type) Output should be a Word or Excel file.
- Create the table in oracle/mysql DBMS and populate sample data to be used in your queries.
- Construct a sample of SQL queries (5 – 8 queries) using your physical model design which can be used to answer possible questions by the decision maker as described in the case-study above. List the business question with each query. Output should be a Word file
- A report of maximum 2 pages is required to elaborate different types of indexes used in Data warehousing and their usage.
B-Business Requirements You can support the business in various ways including and not limited to :
- How does the average yearly balance vary based on the client's job type?
- Is there a relationship between the client's education level and their decision to subscribe?
- Do clients with a personal loan tend to subscribe more or less frequently compared to those without a loan?
- Are there any notable differences in the contact duration for subscribed and non-subscribed clients?
- Develop a sample dynamic dashboard to keep track of the subscription rate “you can add as many KPI’s as you want that serve your task”
- Apply Data cleaning ,exploratory data analysis and feature engineering before fitting the model
- Develop a machine learning model that can be used to classify if certain client will subscribe a term of deposit or not
- Evaluate your model using appropriate metric and discuss why you used this metric
- Discuss which more important the precision of the model or the recall
- Apply hyperparameter tuning on your model and compare between model performance before and after tuning