Data Analytics and Numerical Modeling That Maximizes Productivity
Is Your Data
Helping You Make
the Smartest Decisions?
The Matrix Analytics Group provides expert perspectives based on data analytics and numerical modeling that helps drive optimal decision-making for mining operations.
By combining decades of mining expertise with modern data analytic techniques, Matrix Analytics facilitates discussion that’s aligned with the strategic goals of its clients. Utilizing a model-based approach, Matrix Analytics provides diverse insights into business and operations.
Based in Johannesburg, South Africa, Matrix Analytics’ current focus is hard rock mining in South Africa and select mining operations in the United States.
Traditional Operation Study
- Mine and cost modeling
- Business case modeling
- Scenario analysis
- Flexibility, opportunity cost
- Acquisitions & mergers
- Peer review
- Market research/competitor analysis
- Historical performance
- Statistical analysis
- Simulation / “What if” analysis
- Genetic algorithms
Matrix Analytics FAQ
The Matrix Analytics Group has produced hundreds of reports which have helped our mining and industrial operation clients make smarter decisions. Below are some example questions we have received from them.
Q: Can we identify employees with poor attendance records, likely related to personal problems, in order to manage high risk attendance behavior?
A: Matrix Analytics was asked to analyze the daily attendance of 4 000 employees to identify attendance behavior that is correlated to a higher injury risk profile. We then developed a system to alert the appropriate manager to a change in the risk profile of the employee. The mine had observed that injuries were more frequent in employees whose attendance had recently changed, often due to personal circumstances, and were looking for statistical confirmation of their observation and a means to procedurally identify these changes daily across a large employee base.
Q: Can you create a strategic framework that can quantitatively rank business decisions and options?
A: Matrix Analytics is frequently tasked to develop a process to quantitatively evaluate business decisions and compare strategic options. This process evaluates proposed decisions and options by ranking them by their value generation and strategic fit against the business’s strategic objectives, given a set of scenarios derived from the analysis of market, industry, and socioeconomic experts.
Q: How do we quantify the business’s performance to the market?
A: We analyze the business’s performance, providing important data that is used by management to communicate the company’s interim or annual performance. This includes collaborating with senior executives to construct written and verbal communication.
Q: We track hundreds of KPIs. Can we statistically identify 5-10 key KPIs that explain business performance?
A: The short answer is yes! With today’s connected mines, there is no shortage of KPIs tracked within the business across various business functions, including operating performance, capital project performance, environmental, social and governance responsibilities and human resource development. Matrix Analytics can identifying the most important KPIs that the executive management team and the board need to provide the best explanation of the business’s performance.
Q: Can you run multiple scenario analyses on business cases programmatically?
A: A client asked us to construct and model ten (10) business cases with multiple mines using varying mining methods, treatment solutions and capital timing. Given the number of varying parameters and decision points, the standard process of developing individual business cases within a scheduling program, assigning costs and capital to the production profile, and then conducting a financial evaluation would have significantly increased the project duration and cost. We developed a programmatic solution that could dynamically reschedule, re-estimate Opex and Capex and evaluate profitability for all business cases. This was extended to allow for scenario analysis, which through a Decision Tree, identified the best performing business case across varying pricing and cost scenarios.
Q: Can you calculate cycle times for equipment using data, particularly using non-traditional data that is not directly related to timing?
A: With no available cycle time data and limited budget, a client requested a data model that could estimate cycle times for shuttle cars using engine and conveyor power draw. Using a recurrent neural network, we developed a model which could classify the 100 000 daily data points, per machine, into an equipment activity, which when aggregated allowed for the calculation of the shuttle car cycle times.
Q: How do we optimize our mining/fleet cycle by removing bottlenecks in key processes?
A: We were tasked with reviewing the mining cycle in order to identify the activity that was creating a bottleneck within the process. Cycle times of the fleet were used to create a theory of constraints model, which identifies how the mining process and associated fleet could be improved to de-bottleneck their mining cycle.
Q: Can you identify equipment component behavior that is consistent with component failure?
Q: We were tasked with creating an IoT solution that identifies abnormal operating ranges of various components synonymous with an operating state before a major component failure. The analysis initially identified the statistically significant operating signals from the entire list of tracked signals as inputs into the model. The model then identified the operating states of each of the identified signals that preceded a major failure to create an early warning system. This allowed the client to predictively replace components to prevent a major failure and its resulting unexpected downtime.
Q: Is it possible to identify the number and location of operators underground to assist in dynamically adjust ventilation requirements to reduce power consumption?
A: We constructed a model to continuously track the number of equipment operators underground, based on their clocking data, which allows the ventilation manager to dynamically adjust the in-flow of air aligned to the operating fleet which is measured by proxy by the number of equipment operators underground.
Q: Can you quantify RISK for a project?
A: Matrix Analytics uses multiple models and trials to estimate the impact of various risk events on a project, which is used to estimate the range of possible production, cost and financial outcomes.