Ruben van Gemeren

ML Pipeline for weather forecasting

Automated weather forecasting pipeline for PM2.5 levels in Rotterdam

Machine LearningFeature EngineeringPipeline

Challenge

Automating the creation and updating of a ML predictor model based on sensor data for predicting PM2.5 levels in Rotterdam.

Approach

Creating a pipeline that creates features from sensor data and trains a regression model using PyTorch, which is updated every day to predict the next week's PM2.5 levels.

Impact

Create a webpage that displays the predicted PM2.5 levels in Rotterdam, with a live update every day.

Tech Stack

PythonPyTorchHopsworks