By Dr Ali Soofastaei, CEO, Innovative AI

Many mine sites invest in advanced analytics and automation without a clear view of their current capabilities or a structured road map for progression. This is where a Mining Analytics Maturity Assessment Framework can be designed to support mine-to-mill optimisation and broader value chain transformation.

Mining companies are under sustained pressure to increase productivity, reduce energy consumption and improve safety, while dealing with lower-grade ore bodies and more complex operations. Digital technologies – such as industrial Internet of Things, advanced analytics, digital twins and artificial intelligence (AI)–enabled decision support – are now seen as essential levers for achieving these objectives; however, the success rate of analytics and AI projects in the industry remains mixed. Many initiatives stall because data is fragmented, governance is weak or the organisation is not ready to embed analytical insights into daily decision-making.

A central challenge is that most mine sites lack a systematic way to measure their readiness for digital transformation. Without a clear baseline, it isn’t easy to benchmark performance across sites, justify investment, or design a phased road map from basic reporting towards predictive and prescriptive optimisation.

Maturity models have been widely used in software engineering, data management and manufacturing to structure capability development. In mining, a tailored maturity assessment can provide a powerful lens for understanding how advanced analytics capabilities evolve across the value chain – from drilling and blasting, through to loading, hauling, crushing, grinding, and processing, and to maintenance and reliability.

Digital transformation and analytics maturity in mining

Digital transformation in mining can be viewed as the progressive integration of:

• data infrastructure (sensors, telemetry, historians and cloud platforms)
• analytics and AI (descriptive, diagnostic, predictive and prescriptive models)
• business processes and culture (decision rights, workflows and incentives).

In early stages, digital initiatives focus on basic data collection and reporting: establishing a single source of truth, automating dashboards and improving data quality. As maturity increases, the focus shifts towards real-time monitoring, anomaly detection, predictive maintenance, and, ultimately, closed-loop optimisation of processes such as haulage, comminution or flotation.

Experience from mining and other industrial sectors shows that analytics maturity is multidimensional. It is not enough to ask, ‘Do we have data?’ We must also ask:

• Are our systems integrated and robust enough to support real-time analytics?
• Is our data clean, labelled and contextualised for modelling?
• Do our people have the skills and confidence to use advanced models?
• Are our processes designed to act on analytical insights systematically?

A maturity assessment framework should therefore capture all these aspects, not just the presence of technology.

Mining analytics maturity assessment framework

The proposed framework is designed as a self-assessment tool that can be applied at the site, within a value chain segment (e.g., mine to mill), or in a specific functional area (e.g., mobile fleet management). It consists of:

• six maturity levels describing the evolution of analytics capabilities
• four capability domains (system, data, people and process)
• 28 indicators that can be scored to characterise the current state and identify gaps.

Maturity levels

Table 1 summarises the six maturity levels in the framework. These levels follow a logical progression from basic data collection (sense) to proactive optimisation (predict and act).

Level	Name	Typical capability  and questions answered
1	Raw data	Data is captured, but noisy, incomplete and siloed. Do we have any data?
2	Cleaned data	Data is cleaned, validated and structured. Can we trust our data?
3	Standard reports	Dashboards and reports provide descriptive analytics. What happened?
4	Generic predictive analytics	Basic statistical or machine learning models applied to historical data, often in an offline setting. Why did it happen?
5	Predictive modelling	Asset- or process-specific predictive models in regular use. What will happen?
6	Optimisation	Real-time or near-real-time optimisation, often in a closed loop with control systems. What is the best that could happen?

Capability domains and indicators

The framework evaluates 28 indicators across four domains. Each indicator is scored on a scale from 
0 (not present) to 6 (highly optimised). The individual scores can be visualised as radar plots (spider diagrams) to highlight strengths and weaknesses.

System

Focus on: technical infrastructure and organisational mechanisms that support analytics. The seven indicators are:

1. measurement (existence of KPIs and instrumentation)
2. implementation (deployment of analytics tools into production)
3. innovation (capacity to experiment and scale new ideas)
4. documentation (technical and operational documentation quality)
5. big data platform (scalability and integration of data platforms)
6. collaboration tools (for cross-functional work on analytics)
7. discipline (governance and adherence to standards).

Data

Focus on: quality, accessibility and analytics-readiness of data across the value chain. The seven indicators are:

1. data collection standards
2. reporting standards
3. data formats and interoperability
4. data cleaning and validation practices
5. data storage (historian, cloud and backups)
6. data labelling and contextualisation (e.g., shift, equipment and ore type)
7. data analysis workflows (from exploratory analysis to model input pipelines).

People

Focus on: skills, capacity, and leadership related to data and analytics. The seven indicators are:

1. analytics training for engineers and supervisors
2. years of experience working with data-driven projects
3. number and seniority of analysts/data scientists
4. statistical and data literacy skills
5. modelling and AI/machine learning skills
6. collaboration skills (domain + data science + IT)
7. leadership commitment to data-driven decision-making.

Process

Focus on: how analytics is embedded in operational and strategic processes. The seven indicators are:

1. proactive analysis (continuous improvement versus reactive troubleshooting)
2. modelling practices and life cycle management
3. optimisation practices (use of simulations, scenario analysis and optimisers)
4. documentation of analytical workflows and decisions
5. predictive modelling adoption in routine operations
6. value analysis (linking models to cost, production, risk, and environmental, social and governance metrics)
7. alarm and event reporting (thresholds, escalation paths and root cause analysis).

Domain	Indicator	Example of high maturity (score 5–6)
System	Big data platform	Unified data lake integrating fleet, plant, geology, and maintenance data with governed access and standard application programming interfaces.
Data	Data labelling	All haul and plant records are tagged with ore type, location, shift, operator and circuit configuration.
People	Modelling skills	The in-house team routinely builds and maintains machine learning models to improve throughput, reduce energy intensity, and monitor equipment health.
Process	Value analysis	Each analytics use case has a quantified business case with tracked benefits (e.g., cost per tonne, kilowatt hours per tonne, downtime avoided).

Applying the Framework to Mine Value Chain Optimisation

The actual value of a maturity assessment lies in linking it to concrete operational use cases. In a digital transformation program, the framework can be applied across the mine value chain to:

• prioritise which segments are ready for advanced analytics
• identify where foundational work (data quality, instrumentation and skills) is required
• design realistic and staged road maps for optimisation.

Haulage and mobile fleet

For truck–shovel operations, digital transformation might target:

• fuel consumption optimisation
• payload management and 
  overloading/under-loading reduction
• cycle time and queuing optimisation
• component life and failure prediction

A site at level 2–3 in maturity may have cleaned data, and standard reports on fuel burn and cycle times, but no predictive models. The assessment might reveal:

• adequate telemetry from trucks, but inconsistent labelling of haul segments and ore types (data domain gap)
• little in-house capability for model development (people domain gap)
• no formal process for embedding model outputs into dispatch or maintenance decisions (process domain gap).

The road map would therefore include standardising data labelling, building basic predictive models for fuel consumption and maintenance events, and establishing processes for dispatchers and maintenance planners to act on the predictions.

Comminution and mineral processing

For crushing and grinding, advanced analytics can be used to:

• stabilise the operation of SAG and ball mills
• optimise throughput versus energy consumption versus product size
• predict liner and media wear, and plan change-outs
• link upstream fragmentation to downstream performance.

A plant at level 4–5 might already use generic predictive analytics, but still rely heavily on manual tuning. The maturity assessment may show:

• strong system and data scores (distributed control systems (DCS)/supervisory control and data acquisition (SCADA) integration, and rich sensor data)
• moderate people scores (limited analytics training among metallurgists)
• low process scores (lack of formal optimisation workflows and feedback loops).

Targeted actions could include training plant metallurgists in analytics tools, implementing model-based advisory systems for operators, and moving towards semi-autonomous controls of mill load and grind.

Maintenance and reliability

Across the value chain, predictive maintenance is a common target for digital transformation. The framework helps distinguish between:

• sites that are only monitoring events (levels 2–3)
• sites that are predicting failure probabilities (level 5)
• sites that are optimising maintenance schedules considering production constraints and risk (level 6).

By scoring the maturity of enabling elements – sensor coverage, failure labelling, computerised maintenance management system integration and planner skills – the assessment provides a rational basis for selecting where and how to deploy predictive maintenance solutions.

Implementation road map

A practical way to introduce the maturity assessment into a mining organisation is through a structured but agile approach:

1. Define scope and stakeholders
   Decide whether to focus on a single value chain (e.g., mine-to-mill at one site) or run a benchmarking exercise across multiple operations. Engage operations, maintenance, geology, IT/OT and data teams.
2. Run structured assessment workshops
   Use the 28 indicators as prompts in facilitated workshops. For each indicator, capture evidence and agree on a score from 0 to 6. Technical and operational staff must participate to ensure shared ownership.
3. Visualise current state
   Produce tables summarising scores by domain and level, along with radar plots and maturity curves. These should make it obvious where capability is strong and where it is limiting value creation.
4. Identify priority gaps and opportunities
   Link low-scoring indicators to existing or planned digital initiatives. For example, a plan to deploy a digital twin for grinding may depend on addressing data labelling and skills gaps first.
5. Develop a phased road map
   For each domain, define short-, medium- and long-term actions. Early phases often focus on data foundations and skills; later phases introduce more ambitious predictive and optimisation projects.
6. Integrate with governance and budgeting
   Use the maturity assessment results to support capital allocation and governance decisions. Sites with higher maturity may be selected as pilots for advanced solutions, while others may focus on foundational work.
7. Repeat and refine
   The maturity assessment should be repeated periodically (e.g., annually) to track progress. Thresholds for ‘minimum maturity’ may be defined for deploying certain classes of digital solutions.

Domain	0–12 months (foundation)	1–3 years (scale and integrate)	3+ years (optimise and innovate)
System	Map existing OT/IT landscape (SCADA, DCS, fleet and historians).Establish basic standards for connectivity and data acquisition.Pilot a central data repository (e.g., historian or cloud data lake).	Integrate key value-chain systems (fleet, plant, maintenance and geology) into a unified data platform.Introduce standard application programming interfaces for analytics and reporting.Formalise change management and configuration management for analytics systems.	Implement highly available, scalable big data platforms supporting real-time and batch analytics.Enable digital twins and advanced decision-support tools in production.Adopt continuous delivery pipelines for analytics applications.
Data	Define and document data standards (tags, units and naming conventions).Start systematic data cleaning and quality checks.Identify critical data gaps (sensors, tags and context fields).	Automate data cleaning, validation and enrichment pipelines.Standardise contextual labelling (ore type, location, shift, operator and mode) across key processes.Establish governed data catalogues for analytics users.	Implement active data quality monitoring with alerts and KPIs.Support streaming and event-based architectures for near-real-time analytics.Enable cross-site benchmarking using harmonised data models.
People	Identify key champions in operations, maintenance and processing.Run foundational training in data literacy, statistics and visualisation tools.Clarify roles and responsibilities for analytics projects.	Build a cross-functional analytics team (domain engineers + data scientists + IT/OT).Develop structured training programs for modelling and optimisation.Introduce communities of practice for digital and analytics.	Embed analytics skills into standard technical roles (engineers, planners, supervisors).Establish formal career paths in analytics and digital roles.Promote leaders who consistently use data-driven decision-making.
Process	Select a small number of high-value use cases (e.g., haulage fuel, mill throughput).Document current decision-making workflows for these use cases.Introduce basic KPIs linking analytics to production and cost outcomes.	Embed predictive models into operational workflows (e.g., dispatch rules, maintenance planning, plant set-points).Standardise model life cycle processes (development, validation, deployment, retraining).Create governance forums to review value realisation from analytics.	Implement closed-loop optimisation where appropriate (e.g., APC with machine learning advisers, optimised maintenance schedules).Integrate analytics into strategic planning and budgeting processes.Continuously refine use cases based on performance and emerging technologies.

Conclusions

Digital transformation in mining is not a single project, but is rather a long-term evolution of capabilities. A Mining Analytics Maturity Assessment Framework provides a structured, transparent and repeatable way to understand where an organisation stands today, and how it can progress toward advanced, optimisation-driven operations.

By defining six levels of maturity and assessing 28 indicators across the system, data, people, and process domains, mining companies can:

• benchmark sites and value chains in a consistent way
• align technical and business stakeholders around a common language
• prioritise investments in data platforms, analytics and skills
• reduce the risk of failed digital projects by ensuring readiness
• focus on high-impact use cases that match current maturity, while building towards more advanced capabilities.

When used as part of a broader digital strategy, the maturity assessment becomes more than a diagnostic tool; it becomes an enabler of strategic decision-making, helping miners move systematically from ‘sense and respond’ to ‘predict and act’ across the entire mine value chain. 

End notes

• Soofastaei, A., Davies, J., & Antonio, E. (2018), ‘Development of a maturity scale for mining performance and maintenance data analytics,’ Australian Resources & Investment, 12(1), 10–13. 
• Soofastaei, A. (2016), ‘Development of an advanced data analytics model to improve the energy efficiency of haul trucks in surface mines,’ PhD thesis, The University of Queensland, School of Mechanical and Mining Engineering. 
• Soofastaei, A., & Davies, J. (2016), ‘Advanced Data Analytics: A new competitive advantage to increase energy efficiency in surface mines,’ Australian Resources & Investment, 1(1), 68–69. 
• Gökälp, M. O., Gökälp, E., Kayabay, K., Koçyiğit, A., & Eren, P. E. (2021), ‘Data-driven manufacturing: An assessment model for data science maturity,’ Journal of Manufacturing Systems, 60, 527–546. 
• Halper, F., & Krishnan, K. (2013–2014), ‘TDWI Big Data Maturity Model Guide: Interpreting Your Assessment Score,’ TDWI Research. 
• Russom, P. (2011), ‘Big Data Analytics: TDWI Best Practices Report,’ TDWI Research. 
• TDWI (2013), ‘TDWI Big Data Maturity Model and Assessment Tool,’ TDWI Research. 
• Soofastaei, A. (2023), ‘Analytics Maturity Assessment Solution – Project Proposal for Newmont,’ Internal proposal document.