Big Data Integration

Harness the Full Potential of Census & Other Geo-Based Data

Census data are available to purchase in many countries. These data are relatively affordable (sometimes even free) and represent a valuable source of information about key characteristics on every person in the population. They provide details about potential patient profiles and contribute to explaining how busy medical practices are. In our experience, such data represented at least one third of the total predictive power in all segmentation projects in which we were able to use them.

These data however may seem overwhelming at first sight, as they typically consist of hundreds if not thousands of characteristics available at a very granular geographical level. We offer efficient analytical tools that address both issues and will allow you to make the most of these data.

Dealing with Large Numbers of Characteristics

The number of characteristics in the census is usually quite large. To work around this, a simple option consists of handpicking a reduced number of characteristics. While the selected characteristics are likely to be relevant, the risk associated with this approach is to miss out on several other key characteristics that would play an important role in subsequent analyses. To circumvent this situation, we use a methodology based on data reduction techniques that transforms the original data into a workable (smaller) number of features while maximising the information retained. The approach is able to adapt to different types of data (e.g. # of people, categories, …) and the objective of the study.

Our Catchment Model

When dealing with geo-based data, the key question consists of establishing how each geographical unit impacts a given HCP given his/her practice location. Obviously, the further away the unit the less impact it has. A catchment model is a mathematical tool that formalises the needed computations. Such a model is often called a gravity model, as it computes the attractiveness of a practice location based on the distance of its potential patients.

Using our proprietary catchment (attraction) model, it is possible to determine the number of people with given characteristics within reach of any practice. This can be done with any type of geo-based data and any level of geographical details. In the case of census data, combining this tool with the reduction of characteristics described above allows the generation of an optimised characterisation of the neighbourhood of any practice

Other Large Data Sources

It is possible in many countries to access databases of HCP activities over time. They are either offered by third-party vendors or even publicly available. In this case too, the amount of available information can be overwhelming and a sound strategy to extract meaningful summaries is crucial.

Our team has developed an expertise in this field and can guide you through all steps involved, starting with the data selection according to your goals up to the implementation of adequate analytical tools to compute the most appropriate indices.