KNIME Weather Data Cleaning and Model Training

This workflows reads data from Snowflake database containing sensor readings. Seconds are removed in time stamps and temperature readings are averaged over the hour. Missing timestamps are introduced and filled using Linear Interpolation. Reasonable number of rows are considered for training and testing. Data is partitioned and SARIMA model is trained, finally results are visualized in the component along with original values.

External resources

• An Introduction to Integrated Deployment
• S.A.R.I.M.A. Seasonal ARIMA Models with KNIME