Skip to main content

Train Spotting with Raspberry Pi and Data Science

Harrison Mebane (Silicon Valley Data Science), Stephen O'Sullivan (Data Whisperers)
Open Hardware
E147/148
Average rating: **...
(2.67, 3 ratings)
Slides:   1-PPTX 

How hard can it be to detect a train? In this talk, we discuss our adventures in setting up a system for collection and analysis of visual and audio sensor data in order to detect commuter trains as they pass by our offices in downtown Sunnyvale. Topics covered span from open hardware and sensors to data transfer and storage to machine learning and signal processing.

We will detail experiments with collecting sound data via both Arduino and Raspberry Pi, during which we delved far more deeply into the mechanics of ADCs than ever anticipated. Audio and video data were routed through Apache Flume to HDFS for storage and analysis. Finally, we will explain our use of Fourier transforms and machine learning algorithms in Python, as well as the integration of sensor data with schedule data from Twitter and the Web.

Photo of Harrison Mebane

Harrison Mebane

Silicon Valley Data Science

After several years spent on a PhD in theoretical physics, I recently entered the “big data” community to apply what I’ve learned. I work primarily in Python but am always looking for new tools. I work closely with data scientists and engineers skilled in machine learning and data pipelining.

Photo of Stephen O'Sullivan

Stephen O'Sullivan

Data Whisperers

A leading expert on big data architecture and Hadoop, Stephen brings over 20 years of experience creating scalable, high-availability, data and applications solutions. A veteran of WalmartLabs, Sun and Yahoo!, Stephen leads data architecture and infrastructure.