Areas of Practice
Data Analysis
We provide customised graphics and visualisation for multi-dimensional data that highlight key features of the data. Our strength is in predictive analytics and the formal testing of statistical hypotheses. The data analysis methods we specialise in are: time series analysis and forecasting; spatial data analysis; non-linear mixed effects models; and remotely sensed imagery.
Data Handling
Pink Lake is a firm believer in data governance, where data quality metrics to measure workflow quality and efficiency improvements are applied as part of a continuous improvement cycle. We are well versed in data cleaning, shaping, munging, and scraping as part of the ETL (extract-transform-load) process. Moreover we have led and trained data governance teams to facilitate a change in an organisation's data practice, and to provide positive business cases for proposed improvements in workflow. We can document and model both data and workflows, and provide automated reporting. We also provide R package development.
Data Design
Design begins with measurable and testable performance criteria, from which optimal data capture strategies may be developed. Pink Lake specialises in developing spatially and temporally optimal designs for different contexts: sampling, monitoring, surveillance, survey and experimentation. Our designs have been applied to customer surveys, biosecurity surveillance, population survival analyses, and in the training of machine learning algorithms, among others.
Decision Tools
Data have high value when it supports evidence-based decision making. The outputs of statistical modelling allow measures of risk to be developed, and demonstrate the extent to which management targets have been met. We employ mathematical programming, control charts, benefit-cost analyses, simulation, scenario testing and mathematical modelling to value and explore the consequences of risk-based decision making.
