Data quality is paramount to any research project or decision-making. We identify and correct data entry errors and/or inconsistent values in your datasets, impute missing values, create derived variables from algorithms or project designs, combine different datasets to create an analytical dataset for your analytical models.

We provide survey statistics: instrument design, sampling design (simple random sample, stratified sampling, cluster sampling, multistage sampling); sample size calculation, non- response analysis, data weighting and post-stratification, design effect, margin of error, and survey data analysis.

We analyze your data to find the hidden gem, from experimental design to hypothesis tests to multivariate analyses, including parametric and non-parametric statistics.

We also use data visualization to summarize your data (raw data or processed data) or to present key findings in an easy-to-digest format.

In an advanced simulation analysis, we generate synthetic data to rigorously test your hypotheses and determine the most effective analytical methods for your specific research questions. These virtual observations are designed to accurately mimic real-life parameter values, providing crucial insights into how your outcomes will vary with changes in those parameters.
 In an optimization analysis, we streamline your decision making process to identify the various solutions that yield optimal outcomes under various constraints. We perform sensitivity analysis to show how the outcome changes when the values of the input parameters change.

With machine learning, we use supervised, semi-supervised or unsupervised learning to provide additional insight into your data. With AI we develop programs that automatically learn from your data, structured or not, to identify new patterns to inform your decisions.

When your observations are not really independent, or there are more variables than observations, or there are prior knowledge that you want to incorporate into your models, we use advanced statistics to answer your research question: they include multilevel/mixed models, partial least squares (PLS) models, Bayesian models, time-censored (survival) models, item-response theory models, etc.