Drug Price Prediction

By utilizing computer vision technologies, we can extract information regarding in-store habits. By aggregating this with campaign data, we can generate high-value personalized studies.

Objectives

• Extraction of customer movements, pauses, and habits at the shelves.
• Traffic and detection heatmaps.
• Correlation of information with offers and campaigns.

In today’s demanding landscape, where companies require both flexibility and results, it is essential to maintain exhaustive control over market prices, preventing erosion and anticipating fluctuations to maximize key sales moments. An indispensable tool today is the use of internal, third-party, and environmental data—factors we cannot always control, but whose monitoring and proper processing allow us to stay ahead, becoming faster and more precise.

What we do

We develop a well-organized and classified dataset, enabling data scientists to extract high-value insights that improve company results. This process uncovers potentially unknown relationships and behaviors between medications, as well as connections with internal and external variables that, together, allow us to predict future behaviors.

How we do it

Variable Selection and Clustering

We collect both internal and external variables with a potential relationship to drug prices, combining expert knowledge with the use of regularization models. A clustering process may be necessary to discover twin behaviors.

Predictive Model

Historical drug price information allows us to train artificial intelligence algorithms with these selected variables, aiming to predict both the evolution of current medications and the price of new drugs on the market.

Visualization and Exploitation of Results

We complement the dashboards from the previous phase with new indicators, specifically the results obtained in the classification process and price evolution predictions, within a UX-sensitive environment.