If you are a business that is looking to expand, where should you target your marketing efforts? One way to answer this question is by evaluating the profile of your existing customer base, and identifying areas whose residents or workers share a similar profile.
By mining data from the existing customer database of a local business, and combining this with other metrics (e.g., proxies for purchasing power, land-use policy and education), we developed a model that predicted: (1) expansion zones that have many potential customers but few existing customers, (2) areas that are already ‘maxed-out’ in terms of customers, where maintaining customer satisfaction is key, (3) areas that have more customers than would be expected, suggesting opportunities for targeted marketing, and (4) areas that currently have little potential for expansion. Explore our interactive map (Note: these are all simulated data to protect privacy).
Mental resilience, or the ability to cope with physical or psychological stresses, is an important determinant of professional success as well as personal quality of life. Therefore, quantifying and monitoring mental resilience has important implications for the wellbeing of individuals (e.g., prevention of suicides), as well as for employers and society as a whole (e.g., by reducing costs associated with treatment). It also has important implications for human resource management.
High Level Analytics is working with Professor Ibolja Cernak from the Canadian Military and Veterans’ Rehabilitation Research Program at the University of Alberta to study resilience in Canadian soldiers using data from the Resilience Enhancement in Military Populations through Multiple Health Status Assessments (REIM) project.
We intend to apply these insights to other fields such as emergency services, construction, and corporate human resources.
Wearable devices require complex analytics running in the background to extract meaningful patterns from the surrounding noise. For a new prototype of a heart pulse device, we examined methods to reliably compute several technical metrics (such as maxima, minima, steepest slopes, and dicrotic notch). Our signal exploration toolset was designed to create data frames of metrics which will allow statistics on larger data sets. The main algorithm running in the background is a learning state-machine which allows the program to work on real time or static data.
Feature extraction of a two (red and blue) pulse signal. The time difference between maxima is illustrated in green.