Notifications are important for the user experience in mobile apps and can influence their engagement. However, too many notifications can be disruptive for users. In this work, we study a novel centralized approach for notification optimization, where we view the opportunities to send user notifications as items and types of notifications as buyers in an auction market.
The full dataset, instagram_notification_auction_base_dataset.csv, contains all generated notifications for a subset of Instagram users across four notification types within a certain time window. Each entry of the dataset represents one generated notification. For each generated notification, we include some information related to the notification as well as information related to the auctions performed to determine if the generated notification can be sent to users. See the README file for detailed column decriptions. The dataset was collected during an A/B test where we compare the performance of the first-price auction system with that of the second-price auction system.
The two derived datasets can be useful to study fair online allocation and Fisher market equilibrium. See the README for details and a link to the scripts that generate the derived datasets.
Title
Fair Notification Optimization: An Auction Approach
Issued Date
2023-09-15
Version
v1.1
Status
Published
Citation
Kroer, Christian, Sinha, Deeksha, Zhang, Xuan, Cheng, Shiwen, and Zhou, Ziyu. Fair Notification Optimization: An Auction Approach. Inter-university Consortium for Political and Social Research [distributor], 2023-09-15. https://doi.org/10.3886/k3qx-xz33
Time Period
2022, 2023
Platform
Instagram
Data Formats
web platform data
Software Description
Implementation of algorithms for auction-based notifications and code to generate derived datasets appearing in paper "Fair Notification Optimization: An Auction Approach"
Programming Language
Python
Julia
Environment and Dependencies
Jupyter Notebook
Related Resources
Fair Notification Optimization: An Auction Approach, Journal Article, Is Derived From, arXiv, https://arxiv.org/abs/2302.04835