We are kicking off the 2021 Data Science for Society Seminar Series. This year promises to be another set of great speakers that cover research, practice and policy related to Data Science within our society. We look forward to having you at our seminar series.
The dates of the series as well as some of the speakers are now available below. If you want to be updated on the seminar series, please sign up to the DS@UP Newsletter. You can also access our 2020 archive here.
One of the advantages brought by Big Data and Machine Learning to the banking industry is the possibility of producing propensity models, capable of estimating how likely a customer is to acquire a given product. These propensities can be used, among other functionalities, to support cross-sell campaigns. One of the main problems associated with these models is to establish a feedback loop from marketing channels. Customers might be reached through multiple channels, and in addition to that, take some time between being reached and taking up the product, which makes it difficult to know which channel works best and if it was indeed that channel which led to the take-up. The current talk will focus on both problems, i.e. the technical aspects of propensity models and the feedback problem, along with some ideas for future work on this space.
Felipe Melo has been at ABSA for the last 4 years, starting at Barclays Africa in 2018 as a Data Engineer. Currently working on the CVM space developing models for customer engagement and retention, and lately focusing on automation of internal process by applying computer vision technologies for automatic document understanding. Previous experience involves audio processing, recommender systems, Web mining and distributed computing.
The seminars are hosted by the Data Science for Social Impact Research Group, in the Department of Computer Science at the University of Pretoria.