Description
PyData Chicago 2016
Slides: https://cat.app.box.com/s/1c4mvt8eayb5o7g2wp8nsdwbguhgersm
Identifying predominant driving-style patterns in logged time series data of Caterpillar machines is daunting due to the nature and size of the data. However, insight gained from field data can deliver optimized powertrain control software and better machine performance. A solution for finding patterns was built using engineered features, dimensionality reduction, and unsupervised learning.