SpaceTime advanced analytics technologies

SpaceTime Advanced Analytics Technologies

Improve business outcomes with comprehensive, accurate analytics

Get more business value with advanced analytics

Nokia SpaceTime advanced analytics consider complex events and constraints in real time using techniques such as machine learning, streaming analytics and unique correlation. They deliver high-value, actionable insights that enable enterprises and governments to continuously predict asset failure times and optimize their operations.

Descriptive analytics

Discover what happened for a set of predefined periods or other attributes. Review historical performance and understand where operational improvements are required.

Diagnostic analytics

Establish why, when and where something happened. Determine the root cause of an operational failure or unexpected event and identify effective remediation options.

Predictive analytics

Predict a probable outcome based on past behavior and likely future operating conditions. Then use this prediction as the basis for taking preventative measures, estimating remaining useful life and reducing the impact of unplanned events.

Prescriptive analytics

Identify specific actions that will optimize processes, routes, schedules or plans. Use these actions to improve planning in uncertain conditions, schedule and position resources effectively, or determine the most cost-effective routes.

A powerful and flexible analytics portfolio

Our SpaceTime portfolio enables you to use the market’s most complete set of advanced analytics capabilities, including:

Analytics delivery as a core component of our platform and at the edge of the network
A high-performance platform that delivers answers, alarms and insights through fast, accurate and comprehensive analysis of large, complex data sets and streaming data
Distributed libraries of robust algorithms and models that can be applied to a wide range of use cases
Spatial, temporal and nodal data analysis that provides an in-depth understanding of what, where, when, why and how something has happened, is happening or might happen
Seamless integration with diverse data sources and a rich set of visualization formats, which ensures that insights are presented in the simplest and most actionable way
Operationalized analytics interfaces that enable quick action by allowing users to change input parameters on the fly or perform what-if analysis without relying on specialized personnel

Analytics for any application

We offer an ever-expanding library of analytics created by our team of skilled data scientists with resources and support from the world-renowned researchers at Nokia Bell Labs.

Machine learning

Machine learning techniques enable our algorithms and solutions to produce smarter results, answers, insights, predictions and recommendations as they analyze new data.

Streaming analytics

Streaming analytics analyze data in motion to detect specific anomalies and events. They generate alarms and initiate automated responses in real time.

Optimization under uncertainty

Optimization under uncertainty makes decisions in environments where conditions are dynamic and uncertain.

Anomaly detection

Anomaly detection supports a variety of applications, including identifying and detecting asset failures, fraud, theft, cyber security intrusions, process problems and unwanted activity. It is also useful for assessing and improving data quality.

Geospatial analysis

Geospatial analysis reveals insights about the location, concentration and variations of selected and related activities, assets and situations. Results are typically presented using density maps or heat maps.

Temporal analysis

Time series, or temporal, analysis generates insights about how a situation evolved or is likely to evolve over time. It helps determine the root causes of problems and failures and predict future occurrences.

Video analytics

Real-time analysis of video streams can detect anomalous vectors and identify objects. Video analytics uses edge computing resources to identify relevant events and deliver the applicable video clips to operators. This reduces bandwidth requirements and makes it possible for human operators to manage many simultaneous streams.

Powerful algorithms that maximize operational efficiency

Our machine learning applications help asset-intensive industries save money and boost return on investment by accurately predicting and optimizing asset health and related operations. They provide insights, recommend actions and make decisions that are too complex for people to make in a timely manner, if at all. With advanced analytics powered by our proprietary machine learning library, you can:

Predict the probable time of asset failure without loss of confidence
“See over the hill” and understand your chance of surviving a catastrophic event, where other analytics solutions may project failure
Optimize maintenance activities such as parts procurement, crew scheduling and down times
Extend asset life to increase your return on invested capital
Reduce maintenance and operating costs

Cutting-edge machine learning delivers results

Machine learning is a type of artificial intelligence that can improve its models without explicit programming. It excels when the number of variants or unknown variables make it hard to accurately project which combination of factors may affect an outcome. Our machine learning models combine sophisticated algorithms to deliver unprecedented insight into your assets and operations.

Unsupervised learning

Unsupervised learning uses algorithms to find hidden structures and relationships in unlabeled data, without knowing what the possible outcomes could be. It’s useful for finding clusters of similar items, detecting whether assets are behaving in an anomalous way and performing multivariate analysis.

Reinforcement learning

Reinforcement learning models learn from the outcomes of the actions taken to maximize future rewards over a given time period.

Supervised learning

Supervised learning creates a model where the outcomes are known. The system is trained using real or example data sets to create a model from which it can respond appropriately to new data. For example, you can use supervised machine learning to predict asset failure times based on historical failure data.

From innovation to insights

Our world-class data science experts have created machine learning approaches that can uncover elusive insights in your data. When you combine these approaches with other models, you can be highly confident in their predicted outcomes or prescribed results.

Bayesian networks

Our flexible probabilistic models represent the seemingly random variables and related dependencies that can lead to asset failure. They can be extended to address multiple asset types in a wide variety of industries.

Optimization under uncertainty

Uncertainty makes finding optimal decisions impossible for humans. By using probabilities instead of definitive values for unpredictable variables our applications find optimal solutions that allow for uncertainty.

Bayesian latent state models

Our Bayesian latent state models capture a rich family of probability distributions that more accurately model real-world phenomena. They create a more complex and highly realistic probability distribution function by allowing a sequence of possibly nonstationary mixture models that blend many simple distributions together. The result is a powerful representation of time series data for analysis.