77 Hands-on: Making Sense of big Data, Machine Learning, and Modeling

Abstract

Abstract Advancements in precision livestock technology have resulted in an unprecedented amount of data being collected on individual animals. Often processing of these datasets can be time consuming, tedious, and prone to human-error if processed with conventional software. Open-source statistical software (e.g., R of Python) can provide users with tools to automate many data processing steps for compiling and aggregating data. However, the steps from data collection to processing and training machine learning (ML) models can be time intensive for those new to statistical programming, with few examples pertaining to livestock. The objectives of this hands-on training are: 1) introduce workshop participants to methods for streamlining data processing tasks in Python and R, 2) demonstrate and provide examples of compiling large accelerometer datasets for determining daily livestock behavior; 3) introduce a suite of classification algorithms and validation testing approaches for classifying accelerometer training datasets, and 4) utilize model predictions to estimate and analyze daily behavior for beef cattle. Real life example datasets and code will be provided to workshop attendees to demonstrate how to take raw accelerometer datasets through a finished machine learning analysis. An example of estimating daily energy expenditure for individual animals using behavior data will be provided to highlight linkages to potential rangeland nutrition modeling applications. To obtain maximum benefit from this workshop, participants should bring a portable laptop computer to the workshop and will be encouraged to load software and preview content from a shared cloud directory prior to this training.