Improving long-term sales forecasts in the drug industry: a renovation of the lifecycle project
In pharmaceutical companies, accurate drug sales forecasting is crucial as it allows better budgeting and resource allocation, ensuring that financial resources are used efficiently, with the ultimate goal of being able to reach more patients. At the company Novartis, a Data Science Sales Forecast p...
| Author: | |
|---|---|
| Format: | master thesis |
| Publication Date: | 2025 |
| Country: | España |
| Institution: | Universitat Politècnica de Catalunya (UPC) |
| Repository: | UPCommons. Portal del coneixement obert de la UPC |
| Language: | English |
| OAI Identifier: | oai:upcommons.upc.edu:2117/449243 |
| Online Access: | https://hdl.handle.net/2117/449243 |
| Access Level: | Embargoed access |
| Keyword: | Women pharmacists Product life cycle Time-series analysis Machine learning Sèries temporals Predicció de vendes Anàlisi de supervivència Algoritmes d'IA Anàlisi del rendiment dels models Time series Sales forecasting Survival analysis AI algorithms Models performance analysis Farmacèutiques Cicle de vida del producte Sèries temporals--Anàlisi Aprenentatge automàtic Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic |
| Summary: | In pharmaceutical companies, accurate drug sales forecasting is crucial as it allows better budgeting and resource allocation, ensuring that financial resources are used efficiently, with the ultimate goal of being able to reach more patients. At the company Novartis, a Data Science Sales Forecast project which started in 2019 has been constantly evolving, in order to achieve more accurate, stable and explainable sales predictions. This master thesis conducts a complete renovation of the 'Lifecycle' component of the project, which focuses on the prediction of the stage change from the Growth to the Maturity phase of the drug sales curve, ensuring a consistent and high quality forecast for long-term predictions. It proposes and implements a new core modeling approach, enhances the training dataset creation and introduces a new algorithm to refine predictions for the high investment drugs of Novartis. Extensive analysis are conducted in order to justify the changes applied, demonstrating significant improvements in the stability and the consistency of the predictions, being a key factor in the adoption and trustiness of the project for the financial planners. |
|---|