Sandra Nguyen

Industrial Wastewater Details

Leveraging advanced analytics presents a significant opportunity for the proactive identification of non-compliant Industrial Wastewater (IWW) prior to its release into the Athabasca River. The primary objective of the IWW project is to achieve real-time forecasting of fish mortality associated with process water. To achieve this goal, we diligently collected historical data from various sources and leveraged it to develop a sophisticated predictive model with a dataset spanning the years 2010 to 2020. We utilized the Gradient Boosting Regressor, incorporating six key variables as inputs. Furthermore, to enhance the robustness of our predictions, we have integrated 5% and 95% quantile regression techniques, thereby establishing a reliable prediction interval. Additionally, we have implemented a comprehensive dashboard to facilitate the visualization and analysis of pertinent trends.

Technology/Programming languages used: Python, Gradient Boosting Regressor, Databricks, Data Lake Blob Storage, SQL, Power BI