Services

Strategic problem formulation

• Effective data science starts with a thorough understanding of the business problem.
• We can help think through problems at a high level, and provide expert advice on matching appropriate statistical and machine learning methods to your application.

Training

Short courses are available for

• Elementary statistics, including descriptive statistics and the basics of interval estimation and hypothesis testing. Regression modeling, including logistic regression and other generalized linear models.
• Bayesian computation in C++, R, and Stan.
• Bayesian modeling, including hierarchical models for shrinkage estimation, mixture models for clustering and nonparametric modeling, and hidden Markov models for nonparametric time series.
• Time series modeling with Bayesian structural time series models.
• High dimensional regression and classification using sparsity priors.

Data analysis

We can handle virtually any of your data modeling needs. We have ready made solutions available for

• Forecasting
• Predictive modeling
• Clustering
• Causal inference
• A/B testing

Custom software

Bayesian software is dominated by one-size fits all packages like Stan, WinBugs, and JAGS. These packages are great for quickly profiling new solutions, but they can often be sped up substantially by a custom coded algorithm. Bayesian Statistical Solutions can build you a custom R or Python interface to fast C++ code that fits a custom statistical model designed specifically for your problem.