Spatio-temporal forecasting
Our app performs spatio-temporal forecasting of total consumption and clients at analytical region level and forecast horizons.
This is done using world-class spatio-temporal statistical models with adaptive diffusion curves, and state-space predictive representations using georeferenced data from our clients, and public available regional information. Our customized app is very user friendly displaying interactive maps and time series plots for each analytical region.
New product forecasts
Our app performs forecasting for new products in the market as well as their level of adoption uncertainty. Our algorithms are based on adaptive epidemiological diffusion models that automatically update as new sample information is collected overtime.
We achieve this by combining prior information from analogous products in the market and sample information, to observe the evolution of a new product in the market. The mix of these two sources of information is done using probabilistic consistent methods based on the Bayes’ rule, which requires high performance cloud computing.
Taking into account forecast uncertainty helps our customers do financial planning based on different scenarios: optimist, neutral and pessimist, and as a consequence, to analyze financial viability, scenario production, distribution and marketing while taking into account uncertainty.
People analytics
Our app performs predictive analysis to predict which candidates are the most suitable for a position. It also identifies which variables are the key drivers for successful candidates based on relevant metrics for our clients.
This analysis is a major concern for most large organizations as the hiring process implies many resources and choosing the right candidates has very important implications for the company.
Our descriptive algorithms help to visualize the relevant associations between key drivers of the process. Our predictive algorithms identify the candidates that have the highest probability of being task successful.
Our causal algorithms identify the causal relationships between successful candidates and key drivers so human resources can potentially drive the final outcome. In addition our prescriptive algorithms can determine the optimal levels of key drivers to maximize candidates profitability.
Crime analytics
Our app takes different sources of information into account to display interactive maps that guide security policy decisions to decrease criminality.
This is done using world-class business intelligence spatio-temporal visualization algorithms. For instance, displaying dynamic hot-spots maps. It also identifies associations between different variables to uncover potential patterns that can drive security interventions, reducing criminality.
Impact evaluation
We develop apps to perform impact evaluation of different interventions. Our clients can then perform impact evaluation dynamically, rather than hire a one time analysis.
We use the highest academic standards to perform impact evaluation based on machine learning algorithms such as random forest, k nearest neighbors, deep learning, etc. We are also able to apply standard impact evaluation approaches such as difference-in-difference, instrumental variable, regression discontinuity design, etc. This analysis can help to optimize private and public resources invested in social programs.
