A consultant's firm client wants to mitigate its costs on AWS platform and, so, we are in charge of finding a solution for this issue based on 2 solutions:
-
Forecasting Cloud Spend for the Next Quarter: The client wants to develop a forecast for their cloud spend to build their budget for the upcoming quarter. This involves using the cost and billing usage report from the past 18 months to predict the expected cloud spend for the next quarter accurately.
-
Identifying Spend Areas for Optimization Efforts: The client wants to understand which specific AWS services are driving growth in their spend so that they can prioritize optimization efforts. He also wants to know if the efforts on Product and/or on Service should be prioritized and which ones.
We'll decompose the answer of the 2 problems into 4 parts:
- Data Exploration and Analysis
- Feature Engineering
- Forecasting models using ML and DL to answer question 1
- Cluster Analysis to answer question 2
Python script and outputs can be seen at "main.ipynb".