Henriques, Roberto André PereiraRamos, Tiago josé Isidoro2023-02-022023-02-022023-01-23http://hdl.handle.net/10362/148538Internship Report presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data ScienceFood Delivery Digital Platforms have been investigating novel approaches to leverage drivers’ interactions with the platforms to increase operational efficiency. At iFood, one such approach is the introduction of a Booking system, wherein drivers can get increased order priority for scheduling shifts in advance. The theoretical assumption is that this system can increase visibility of drivers’ intentions as well as influence their choice of working regimes, leading to an increase in supply predictability. The present internship report details the first steps towards the discovery and validation of the aforementioned assumption, by applying statistical inference tests to the relevant data, as well as training and testing predictive modeling that leverages the new information available by the system and comparing it to current operational models used internally. The results show the newly introduced variables are important, albeit the quantifiable impact is comparatively small. However, the applied models achieve better performance and test scores than the current internal models. Based on the findings of the project and observational insights from previous studies, future steps are proposed, which include the refinement of current operational models, and business initiatives with a positive impact on the importance and quality of specific variables related to the booking system.engFood Delivery PlatformsGig EconomyDriver Shift Bookingmaster thesis203211626