Menu

Sabisu Analytics

From performance metrics to deep learning with industrial and enterprise data.

Analytics with Sabisu

Meeting all asset, project and portfolio performance requirements

Sabisu combines deep learning and statistical techniques to provide state of the art monitoring and automatically initiate workflows.

All KPIs and metrics are automated using Sabisu Pipelines, aggregating raw data from enterprise and process sources to ensure accuracy and ad hoc analysis at all levels; process, plant, site, project and enterprise.

Key metrics such as time-on-tools are continuously maintained for instant comparison, analysis and optimisation – they’re up to date, all the time.

Sabisu uses deep learning & statistical algorithms in both asset operation and project domains, maintaining reference models (‘digital twins’) which predict behaviour, provide early warning and detect trending risks.

Sabisu integrates process expertise, vendor, EPC and sub-contractor data into analytics and predictive models to improve accuracy.

Project Analytics

Strengthening project controls & providing early warning

Traditional metrics such as Earned Value and Earned Schedule are instantly available, along with progress metrics derived from Primavera/Project, or MS Excel.

Sabisu develops Asset Models throughout the lifecycle to reflect project progress which are then compared to reference models to predict outcome. This also provides an ‘as built’ model which can be passed into operations through CSU.

Reference Models are used to inform early stage forecasts and scope optimisation by acting as a ‘solution library’, reducing engineering costs and linking 6D CAD, vendor and historical performance data to improve predictions.

Modelling previous project performance ensures that previous lessons learned are reflected in current/new projects.

Field progress capture links directly to Sabisu to automatically update schedules & automatic risk analysis, providing early warning during the Execution phase.

Projects

Operational Analytics

Look beyond business intelligence to deliver real control

Sabisu analytics emphasise real-time performance, automating normally manual metrics including variable/fixed costs, quality & maintenance.

Critical Asset Monitoring in Sabisu is the most accurate yet seen, combining statistical and machine / deep learning methods to avoid false positives and noise induced errors.

Asset models reflect performance & link to engineering, maintenance and vendor data to inform forecasts and performance comparisons.

Sentiment analysis of shift and maintenance logs detects repeated asset or process issues which have not been highlighted in other ways.

Field capture links directly to Sabisu to automatically update schedules, reports, analytics, Site/Asset Models and predictions.

Operations

Techniques & Technology

Sabisu has the technology needed real-time control of complex projects & operations

Sabisu uses cloud computing and Spark, a highly scalable and distributed in-memory MapReduce system which is exceptionally fast, wherever possible.

Analytics are executed on a micro-service architecture, ensuring performance at scale. High performance libraries including MXNET and SparkML are used for machine / deep learning.

Customers requiring on-premise deployment have options including integration with Sabisu Cloud Central Services through a secure RESTful web interface or VPN, or implementing a local analytics service.

Artificial intelligence techniques are used to perform sentiment analysis, extracting quantitative information from commentary, narrative & logs which then informs other analytics.

Python is used both in prototyping and production to ensure rapid deployment of custom analytics solutions.

Third party Python libraries are often provided for inclusion in the platform.

More Sabisu Technology