Repository for our project in Experts in Teamwork (EiT) at NTNU. The web app serves predictions for surface water temperatures at future dates, based on historical data from YR / Norwegian Meteorological Institute. Predictions are served via a pretrained Keras model that is loaded in a firebase function with Tensorflow.js.

The ML model consists of time-series data fed through a single convolutional layer which is then fed into two bidirectional layers of LSTM. This structure is commonly known as the "CNN-LSTM" model.

Live demo

• Anders Hoel
• Herman Sletmoen
• Nora Høva Erevik
• Kaja Sofie Lundgaard
• Henrik Lia
• Magnus Schjølberg