Data Analysis and Programming

In increasingly complex global markets, companies face the absolute necessity of properly organizing and analyzing large amounts of data. We have hands-on experience dealing with many different data sources. To gain powerful insights, it is important to identify patterns and relationships between variables through a variety of data mining techniques (e.g. clustering, association).

We use SAS® for statistical computing. For computationally-intensive methods, C/C++ is used. Additionally, PHP programming is used for web applications and text parsing.

Statistical Analysis and Predictive Modeling

We employ advanced statistical techniques to provide consistently effective solutions. The statistical analyses we provide include, but are not limited to:

• Regression Analysis
• Survival Analysis
• Decision Trees
• Multivariate Techniques
• Nonparametric Statistics
• Categorical Data Analysis
• Analysis of Variance (ANOVA)
• Conjoint Analysis
• Significance Testing

Time Series Analysis and Forecasting

In business, there are many factors that change over time: sales, prices, investment returns, etc. It is important to understand the components of a time series such as trend and seasonality. In time series models, historical patterns in a series are examined to forecast future behavior. In explanatory models, we examine how causal variables (e.g. prices, promotions, macroeconomic factors) affect the variable we are trying to forecast (e.g. sales, rates of return).

Monte Carlo Simulation

Monte Carlo methods have a rich historical background. Modern Monte Carlo methods date back to physics research in the 1940's. The main idea is to use random samples to estimate the parameters of a system. Monte Carlo methods have been popular not only in science, but also in finance and statistical inference.

Most real-world systems are subject to uncertainties. Monte Carlo methods allow us to incorporate these uncertainties into our models. Examples of these uncertainties (stochastic parameters) include demand, costs, or interest rates. Monte Carlo methods provide accurate estimates and, unlike other methods, allow testing the sensitivity of model parameters.