▼ Scroll to Site ▼

Newsletters




Understanding Complex Data with Time Series Analytics


Time series analytics can provide game-changing insights for the business, allowing organizations to forecast and predict the future.

Extracting high-quality insight from time series data can be challenging because it is multi-dimensional, complex, huge in scale, and often driven by the Internet of Things (IoT).

DBTA recently held a webinar featuring John O'Brien, CEO and principal advisor, Radiant Advisors, and Eva Nahari, director of product management, Cloudera Data Warehouse, who discussed how to maintain and what to do with time series analytics.

Time series data is different from traditional statistical data, Nahari said. It’s complicated, high dimensional, and non-stationary.

In order to see the value of time series analytics, it’s important to understand the difference between events versus state, O’Brien said.

Events are the series of actions that lead to a stateful outcome, whereas the state is the current status or values of a process.

For example:

  • A chess board of pieces is “state of the game”
  • The outcome of interest = who won
  • Analyzing events of moves = how they won

The outcomes of interest are the result of many conditions that influenced the event, O’Brien said.

With this information, there are several ways to go about analyzing the data including conducting process optimization, prediction, supporting increasing data volumes, and moving beyond normal management with machine learning and automation.

Data analytics platform requirements include:

  • Data ingestion such as continuous data ingestion and continuous data processing
  • Distributed processing with a modern data warehouse enabled with self-service data analytics and machine learning deployments
  • Cloud platforms that can include hybrid and multi-cloud architecture or data enrichment for increased analytics capabilities

“More data streams lead to a better understanding of the conditions and external factors that caused an outcome,” O’Brien said. “Proficiency with time-series data analytics can drive higher quality analysis and more accurate predictions.”

Because existing systems limits reduce time series quality, Nahari recommended the Cloudera Data Platform. The solution offers multi-function analytics, hybrid and multi-cloud, is secure and governed, and is an open platform.

An archived on-demand replay of this webinar is available here.


Sponsors