Measuring Lake Recreation in Iowa: What Mobility Data Add

By Yongjie Ji and Dimitris Friesen

For more than two decades, Iowa State University’s Center for Agricultural and Rural Development (CARD) has conducted a series of water-based outdoor recreation surveys that have contributed to both academic research and policy discussions. These surveys, including the Iowa Lakes Project, have provided detailed and widely used information on in-state lake visitation, recreational behavior, and user characteristics, supporting analyses of water quality, recreation demand, and benefit-cost evaluation. Findings from this body of work have informed state and local decision-making related to lake restoration, nutrient reduction strategies, and outdoor recreation investments. Building on this established survey foundation, recent advances in anonymized mobility data offer a complementary way to measure lake recreation, particularly by providing broader spatial coverage and higher-temporal frequency observations. This report examines what mobility data add to the measurement of lake recreation in Iowa. Readers interested in learning more about the Iowa Lakes Project and related research can find additional information on the project’s public website (https://lakes.card.iastate.edu/). 

The mobility data used in this report come from Cuebiq (Spectus), which compiles anonymized location signals from mobile devices whose users have consented to data sharing through third-party applications. Cuebiq researchers developed an in-house algorithm to decide the “stop” of devices. We then use the spatial relationship between these stops and lake boundaries to define a visit to any lake.¹ We aggregate and process device-level data to protect individual privacy, with home locations and movement patterns reported at coarse spatial scales, typically at the census-block group level. For this analysis, we use mobility observations covering the 2019–2021 period to construct measures of lake visitation that can be compared with survey-based estimates. We define a lake visit as an instance in which a device records a dwell time of more than 30 minutes within a defined buffer zone surrounding approximately 130 lakes included in the household survey.

In comparing mobility data with survey-based measures, we focus on three dimensions that are particularly relevant for recreation analysis: spatial coverage, temporal coverage, and repeatability over time. Specifically, we examine how broad mobility data capture lake visitation across the state, how monthly and seasonal patterns compare with survey-based estimates, and how consistently visitation can be monitored across years. By organizing the comparison around these dimensions, we aim to clarify where mobility data provide additional insight and how they can complement established survey approaches in measuring lake recreation in Iowa.

Unprecedented spatial coverage

The Iowa Lakes Project typically distributes several thousand household surveys in each round to a representative sample of Iowa residents aged 18 and above. In 2019, approximately 5,000 surveys were mailed, and just over 2,000 completed responses were returned. The left panel of figure 1 illustrates the coverage of census block groups (based on the 2010 Census) represented in the survey data. As expected, responses are concentrated around major population centers, such as the Des Moines metropolitan area, while coverage in more rural regions is comparatively sparse. In contrast, the mobility dataset covers approximately 1 million devices during the 2019–2021 period, representing roughly one-third of Iowa’s population. As a result, spatial coverage is nearly universal, with only a small number of census block groups unrepresented. Because these data are collected passively and at scale, this broad coverage enables researchers to observe recreational use across hundreds of lakes simultaneously, including both well-known destination lakes and smaller lakes with lower public visibility.

While mobility data substantially expand spatial coverage, it is also important to examine whether the underlying travel behavior patterns remain comparable to those observed in survey data. To assess this, table 1 reports the distribution of visits across travel-distance bins for both the 2019 Iowa Lakes Project survey and the mobility data. Comparing these distributions allows us to evaluate whether the broader geographic representation in mobility data meaningfully alters the observed distance-decay pattern in lake visitation.

In both data sources, the largest share of trips falls within the 10–30 miles from home range. The survey reports a relatively higher proportion of long-distance trips—particularly in the 90+ mile category—while the mobility data capture a greater share of short-distance visits (within 10 miles). The higher share of long-distance travel in the survey is not surprising, as the Iowa Great Lakes—located in Dickinson County in northwestern Iowa, which are distant from major population centers such as the Des Moines metropolitan area—attract visitors willing to travel substantial distances. At the same time, the broader spatial coverage of mobility data likely improves measurement of shorter, local trips that may be underrepresented in household surveys. Despite these differences in levels, the overall distance-decay pattern remains consistent: visitation declines as distance increases, and mid-range trips remain the dominant category. Moreover, the stability of the mobility-based distributions across 2019–2021 suggests that expanded geographic coverage does not fundamentally alter the underlying structure of travel behavior.

Year-round and high-frequency temporal coverage

A second strength of mobility data is their high temporal resolution. Unlike surveys, which typically provide annual or seasonal snapshots of recreation behavior, mobility data allow visitation to be tracked at monthly—or even finer—intervals. This enables researchers to examine seasonal patterns, short-term fluctuations, and responses to sudden external shocks with greater precision.

The top panel in figure 2 shows a strong concentration of lake recreation during the summer months. When households were asked which month they make most of their lake trips, July and June received the highest responses, followed closely by August. Participation is lower in May and September and drops further in October and other months (November–March), highlighting the pronounced seasonal nature of lake use in Iowa. This seasonal pattern is consistent with the mobility data for 2019 (bottom panel), which shows visitation rising sharply in late spring, peaking in June and July, and gradually tapering off through early fall. Compared with 2019, we see higher visitation in 2020 and 2021, while the timing of peak use remains largely similar, reinforcing the conclusion that mobility data capture the core seasonal dynamics reflected in survey responses.²

This higher visitation levels observed in 2020 and 2021 are also consistent with broader national trends in outdoor recreation following the COVID-19 pandemic. Reports from the Outdoor Industry Association (OIA) document a substantial increase in participation in outdoor activities during this period, as households shifted toward local, open-air recreation options. The mobility data therefore align not only with survey-based seasonal timing, but also with broader shifts in recreation behavior documented at the national level.

Scalable and repeatable monitoring over time

Mobility data also offer a scalable and cost-effective approach to monitoring recreational use over time. Once access to the data is established, visitation metrics can be updated regularly without the need for repeated survey deployments. Because mobility data are generated from observed device movements rather than predefined site lists, researchers and policymakers can readily expand or modify the set of lakes included in the analysis as priorities evolve. In contrast, traditional household surveys typically identify targeted recreational resources in advance, limiting the ability to incorporate additional sites after implementation. This flexibility supports consistent tracking of long-term trends and enables timely before-and-after evaluations of policy interventions, infrastructure investments, or environmental events—even when those changes were not anticipated at the time of the original data collection.

Limitations and interpretation

Definition of trips

Mobility data define lake visits using the pre-defined “stop” records in the Cuebiq (Spectus) dataset, which identify instances in which a device remains within a defined area for a minimum duration. Using this definition, we record approximately 257,000 lake visits in 2019 among 60,518 visitors/devices, or about 4.25 visits per visitor/device. In contrast, the 2019 survey reports an average of 10.44 visits per trip taker.³ Several factors help explain this difference. First, we restrict mobility-based visits to stops occurring within 100 meters of a lake boundary. Second, we require a dwell time of more than 30 minutes, which excludes shorter visits. Third, the stop-based definition may not capture recreational activities involving continuous movement—such as jogging or cycling around lakes—if they do not meet the dwell-time threshold. Fourth, not all the devices we tracked appeared in the mobility data set all the weeks in a year. Together, these criteria likely lead to a more conservative count of lake visits in the mobility data.⁴

Lack of micro information

To protect privacy, mobility datasets do not include individual-level demographic or socioeconomic information, limiting their usefulness for analyzing heterogeneous recreational responses across different social or economic groups. In addition, home locations are typically reported at coarse spatial scales or otherwise masked, which constrains the precision of origin-destination measures and complicates analyses that rely on detailed travel behavior. As a result, researchers using mobility data must rely on additional assumptions when interpreting observed visitation patterns, particularly when drawing inferences about behavior at the individual or subgroup level. These assumptions are often reasonable but should be made transparent, as they can influence conclusions. 

Taken together, these limitations do not diminish the value of mobility data for measuring aggregate patterns of lake use; rather, they underscore the importance of pairing mobility data with complementary sources, such as surveys, when addressing questions related to equity, distributional impacts, and welfare. More broadly, the increasing availability of mobility data also creates new opportunities for empirical research across a wide range of policy-relevant areas beyond outdoor recreation.

Footnotes

1. Interested readers can reach to author Yongjie Ji for a technical definition of a visit. The official categorization of stops can be found in Xiang et al. (2016).

2. Mobility data record trip activity at the daily level; however, for clarity and comparability with survey-based measures, we present the results aggregated to the monthly level.

3. See table 2 of the online survey report at https://lakes.card.iastate.edu/files/inline-files/iowa_lakes_survey_report_2019.pdf.

4. An alternative counting approach would involve working directly with raw device “ping” data to reconstruct movement paths around lakes. However, implementing such a method would require substantial computational resources and data processing capacity.

References

Xiang, Longgang, Meng Gao, and Tao Wu. 2016. "Extracting stops from noisy trajectories: A Sequence Oriented Clustering Approach." ISPRS International Journal of Geo-Information 5(3):29.

Suggested citation

Ji, Y., and D. Friesen. 2026. “Measuring Lake Recreation in Iowa: What Mobility Data Add.” Agricultural Policy Review, Winter 2026. Center for Agricultural and Rural Development, Iowa State University. https://agpolicyreview.card.iastate.edu/winter-2026/measuring-lake-recreation-iowa-what-mobility-data-add