Caltrain Rider from Silicon Valley Data Science

Many people who live and work in Silicon Valley depend on Caltrain for transportation. And because the Silicon Valley Data Science headquarters are in Mountain View, not far from a station, Caltrain is literally in our own backyard. Our R&D program emphasizes investigating emerging technologies, including Apache Spark and Kafka, and giving back to the community to which we belong. So with Caltrain Rider (iOS, Android), we have been using data science techniques to understand and predict delays in the Caltrain system. It’s a complete data product, comprising a mobile app, predictive backend and deriving input from public sources and a a sensor platform.

We built a view of the Caltrain systems using information from some of our own sensors (cameras and microphones) combined with information from publicly available sources (Twitter, Caltrain API). You might think this would be easy, since Caltrain provides schedules and even a real-time API. But as with all trains, Caltrain sometimes gets behind schedule. Sometimes its API estimates are off, and at other times its API crashes and provides no information. So we’ve been attempting to get a larger view of the Caltrain system from a variety of sources.

CalTrainScreenshot from the iOS Caltrain Rider app

We consume, integrate and analyze a variety of local, distributed, redundant, and possibly unreliable information sources in order to understand the state of the system. Data sources are never completely reliable or consistent across a large system, so this gives us experience producing predictions from messy and possibly erroneous data.

We’re excited to be part of Datapalooza, and talk about what’s involved in creating a complete data product: from R&D, to public launch. We’ll use a strong Q&A format to drive our session. On any one of the days, there’ll be a host leader from SVDS, plus some data scientists and engineers from our teams who worked on the project.

Newsletter

You Might Also Enjoy

James Spyker
James Spyker
2 months ago

Streaming Transformations as Alternatives to ETL

The strategy of extracting, transforming and then loading data (ETL) to create a version of your data optimized for analytics has been around since the 1970s and its challenges are well understood. The time it takes to run an ETL job is dependent on the total data volume so that the time and resource costs rise as an enterprise’s data volume grows. The requirement for analytics databases to be mo... Read More

Seth Dobrin
Seth Dobrin
2 months ago

Non-Obvious Application of Spark™ as a Cloud-Sync Tool

When most people think about Apache Spark™, they think about analytics and machine learning. In my upcoming talk at Spark Summit East, I'll talk about leveraging Spark in conjunction with Kafka, in a hybrid cloud environment, to apply the batch and micro-batch analytic capabilities to transactional data in place of performing traditional ETL. This application of these two open source tools is a no... Read More