The rise of 5G and Internet of Things (IoT) technology has brought the option to collect massive amounts of data. However, there are limitations imposed by bandwidth, storage space, and energy consumption and much work is being done to mitigate and loosen these limitations. However, once a hardware solution has been installed, the limit is set and it can be costly to replace it. When the limit is reached, the way forward is to decrease the volume of data by increasing its value or omitting the data with little to no value. The goal is then to devise clever sampling and transmission schemes for collecting only the most valuable data. Value of Information (VoI) is a measure of how useful a data point is, with three important aspects: is it collected at the right time, is it relevant for the problem at hand, and is it providing novelty that could not be predicted?

This thesis explores the use of side information for optimizing VoI in a wireless IoT network setting. Multiple types of side information are considered in order to optimize the different aspects of VoI. First, graph structures are used to increase the novelty of data transmitted. Second, knowledge of the query arrival process is used to minimize the response time. Third, knowledge of the aggregation functions used to summarize the data is used to maximize the relevance of the data collected.