Explore the potential of IoT through a live demonstration showcasing data collection from a smartphone's accelerometer, storage in Cassandra using a timeseries model, and real-time analysis with Spark. Learn about Cassandra's key features, including last write wins and compaction, and discover how to implement timeseries data storage and querying. Dive into the Spark ecosystem, understanding RDD transformations, actions, and cluster deployment. Gain insights into activity recognition using multiclass classification and decision tree models. Witness the process of collecting training data, extracting features, and implementing a random forest algorithm for predictive modeling in this comprehensive exploration of IoT, timeseries data, and machine learning.
IOT, Timeseries and Prediction with Android, Cassandra and Spark
Overview
Syllabus
Intro
Duchess France
Far far away
My connected objects
Accelerometer
Cassandra Key Facts
Last Write Win
Compaction
Timeseries with Cassandra
Storage model: Disk layout
Query patterns
Analyse data
Spark ecosystem
Spark data sources
RDD transformations and actions
Word count sample
Spark on cluster
Spark key facts
Spark Cassandra Connector
Spark connection setup
From Cassandra to Spark
Spark Streaming
Activity Recognition
Multiclass classification
Decision tree model
Supervised learning
Predictive model
Collecting data
Training data
Timeseries: walking vs Jogging
Features extraction
Compute features
Random forest
Taught by
Devoxx
