Recently, unmanned aerial vehicles (UAVs) have been receiving significant attention due to their wide range of potential application areas. To support UAV use cases with beyond visual line of sight (BVLOS) and autonomous flights, cellular networks can serve as ground connectivity points, and they can provide remote control and payload communication for UAV links. However, there are limited data sets to study the coverage of cellular technologies for UAV flights at different altitudes and develop machine learning (ML) techniques for improving UAV communication and navigation. In this article, we present raw LTE I/Q sample data sets from physical field experiments in the Lake Wheeler farm area of the NSF AERPAW experimentation platform. We fly a UAV that carries a software-defined radio (SDR) at altitudes ranging from 30~m to 110~m and collect raw I/Q samples from an SDR-based LTE base station on the ground operating at 3.51 GHz. We adopt a standard metadata format for reproducing the results from the collected data sets. The post-processing of raw I/Q samples using MATLAB's 4G LTE toolbox is described and various representative results are provided. In the end, we discuss the possible ways that our provided data set, post-processing sample code, and sample experiment code for collecting I/Q measurements and vehicle control can be used by other ML researchers in the future.