Interactive Event Sequence Prediction for Marketing Analysts
Fan Du, Shunan Guo, Sana Ambreen Malik, Eunyee Koh, Sungchul Kim, Zhicheng Liu · 2020
Timestamped event sequences are analyzed to tackle varied problems but have unique challenges in interpretation and analysis. Especially in event sequence prediction, it is difficult to convey the results due to the added uncertainty and complexity introduced by predictive models. In this work, we design and develop ProFlow, a visual analytics system for supporting analysts' workflow of exploring and predicting event sequences. Through an evaluation conducted with four data analysts in a real-world marketing scenario, we discuss the applicability and usefulness of ProFlow as well as its limitations and future directions.