Occupancy Prediction
Nexudus, a global leader in management software for coworking spaces, sought to revolutionize the management of its shared resources worldwide. Its objective was to maximize the profitability of these spaces, which include meeting rooms, events and hot-desks, by implementing dynamic pricing. This innovative approach sought to adjust prices based on real demand, taking advantage of its extensive history of reservations and use in more than 150 countries.
At WhiteBox, we approach the problem by developing a next-generation demand prediction model using deep learning libraries such as TensorFlow. This highly accurate model predicts the occupancy of coworking spaces based on historical data, weather conditions, key dates, and more. Implemented on AWS and managed with tools such as MLFlow and Apache Airflow, this system not only anticipates demand, but also adjusts prices in real time, ensuring maximum profitability at any time.
The final result has transformed the way in which companies that use Nexudus manage their resources. These spaces now enjoy unprecedented optimization in the profitability and occupancy of their shared areas. This dynamic pricing system has made it possible to fill spaces that previously remained empty, offering discounts at times of low demand and maximizing revenues at peaks of high occupancy. Thanks to this project, Nexudus software now stands out as a unique and superior solution in the coworking space management market.
- Forecasting the occupancy of more than 40,000 shared resources in more than 150 different countries.