Introducing Cloud Inference API: uncover insights from large scale, typed time-series data
The code in the above animation can be found here.
Whether businesses are measuring clicks, queries, or sensor readings, they’re often generating time series or event-driven data. Analyzing this data offers businesses the potential to uncover insights in real time, but oftentimes it also means building a learning system that can scale to millions or even billions of data streams. For many businesses, this proves to be prohibitively challenging to design.
Today, we’re announcing the Cloud Inference API to address this need. Cloud Inference API is a simple, highly efficient and scalable system that makes it easier for businesses and developers to quickly gather insights from typed time series datasets. It’s fully integrated with Google Cloud Storage and can handle datasets as large as tens of billions of event records. If you store any time series data in Cloud Storage, you can use the Cloud Inference API to begin generating predictions.
Here's more on the benefits of Cloud Inference API:
Simple. Cloud Inference API offers a simple, tree-like query language with well-defined operators, which makes sending queries and processing responses from the system straightforward. The query language allows you to define “interesting” windows in time over which the correlations are computed. Queries that are selective in the time domain help you find spikes and other interesting features in your data.
Real-time. Through its streaming update interface, Cloud Inference API allows you to perform real-time, context dependent analysis over time series data. This can be used in a variety of applications, from suggesting trending topics to detecting anomalies from sensor data.
Scalable. Cloud Inference API performs flexible correlations across strongly-typed time series, with each row following the same, explicit type format. Because the system is agnostic to the data it works with, it can accommodate very large datasets (trillions of events) and sustain high query load (hundreds of thousands of queries per second).
Many different industries can benefit from time-series data prediction. Retailers can analyze foot traffic to sales conversion rates at brick-and-mortar locations or online. Content providers can use collaborative filtering to offer users high-quality recommendations. And IoT companies can correlate multiple sensor data sources in real-time.
Snap Inc. is one company that’s already been exploring the benefits of Cloud Inference API for Snapchat. "Cloud Inference API promises to replace several custom dataflows with a single system that offers very low latency and real time updates,” says Peter Ciccolo, Software Engineer at Snap. “Moreover, the same data can be used for multiple purposes, making it easy to explore new features with minimal custom code."