Catapult BI consultants engaged with Melbourne Water, and
utilised machine learning techniques to predict the pond turbidity level, which could further lead to possible shortages of Class A water.
“Population growth and climate change have accelerated since
the Millennium Drought in the early 2000s, creating a new normal for our water supply.”- quotes Melbourne Water.
Various factors involved include wind, temperature, and ph. balance, etc. Catapult BI consultant created an end-to-end machine learning pipeline by using multiple AWS services including Sage Maker and S3 Bucket, Lambda, and CloudWatch.
Major tasks of the project involve:
-
Data cleaning and transformation
-
Data exploration to investigate quality issues and correlations within the dataset
-
Feature engineering to derive and integrate data
-
Model training, fine-tuning, and evaluation.