Peer Reviewed Article Hypothesis , Vol. 35, No. 1, 2023 Some Like it Quiet: Using a Heat Map to Study Space Use in a Health Sciences Library Will Deana aResearch and Data Services Librarian, Temple University Libraries, Temple University, Philadelphia, Pennsylvania, will.t.dean@gmail.com Objective: This method is used to identify preferred seating locations within the library. Approach: This method involves counting where patrons are situated in the library at different times of the day, at different times of the year, and using that data to generate heat maps. Data type(s) used: The data consisted of locations and counts of patrons around the library at different time points. Strengths: Provides a granular look at space use in the library. Limitations: It is difficult to know why patrons prefer different sections of the library. Overview As with all libraries, health science libraries want to know how their patrons use their space so changes can be made to help create a welcoming and useful experience. To keep track of library use, our library regularly counts the total number of patrons in the library and the number on each floor throughout the day. There is no record, however, of where on each floor patrons choose to study, read, and congregate. Nor are there any records of how those behaviors vary over the course of an academic school year, a week, or hour by hour. A literature review of space use studies in academic and health sciences libraries was conducted and served as the basis for the work described in this article. Khoo1,3 and Gullikson2 described studying space use in academic libraries using heat maps, and this option was chosen for several reasons. First, it allowed for collecting granular data about how patrons were using the space at a particular time instead of relying on those who are motivated to respond to surveys or participate in focus groups. Additionally, heat maps are easy for the viewer to parse and give immediate information about how different spaces are used. Library staff also wanted to create visualizations of how the library was used as part of a 10-year review of the library’s operations, and the data collection could be conducted using existing staff time and operations. The method described below makes it possible to quickly generate heat map images from any number of time points by loading data and variables into Adobe Photoshop. Since data was collected four times a day, five days a week, over the course of two weeks, we wanted to generate heat maps for each time point – forty times overall – with minimal manual effort. This process allows users to generate a heat map for each of those time points quickly and easily and one can modify the data if needed without any manual image manipulation. Methods for this study were adapted from Khoo1 and Gullikson2 such as recommendations on how to divide up sections of the library space into zones that could be analyzed together. 1 mailto:will.t.dean@gmail.com Peer Reviewed Article Hypothesis , Vol. 35, No. 1, 2023 Example In the fall of 2018, a health sciences library in a large, urban research university conducted a space use study by creating a series of heat maps. The intent was to determine usage levels of different areas of the library daily for two one-week periods at the beginning and end of the semester. Working with library circulation staff, a plan was developed to modify pre-existing patron counts to record the exact patron seating locations during specific times throughout the day. Using the methods described by Khoo,1 each floor of the library was divided into zones based on furniture groupings and separation barriers. To gather more granular data, we decided to follow Gullikson’s2 example and count the occupied seats in each zone as well. Using modified paper floor plans, which detailed the different zones and seats on each floor, we tested the data collection process in late November 2018. Library staff counted patrons by marking each occupied seat in some zones and tallying the total number of patrons in others. After feedback from library staff, the floor plans were modified to improve usability and include orientation instructions. The library’s regular weekday hours of operation are 7:30AM-10PM. Staff counted library patrons four times a day every weekday for one week at 9:30AM, 12:30PM, 3:30PM, and 5:30PM. These data collection times were chosen so the data from different dates could be compared easily and to fit with existing staff counts of patrons. Although our library is open in the evening (and overnight during finals weeks) and on weekends, counts could not be done at those times due to staffing limitations. The first set of data was collected during a week near the end of the semester with an expected high occupancy rate. The second set of data was collected during a traditionally low occupancy week following the winter break. Data was transferred from the paper templates into spreadsheets that mimicked the layout of the floor plans, as outlined in Gullikson.2 Calculations were performed to derive an occupancy percentage based on a hypothetical 100 percent occupancy. To create the heat maps, we used Adobe Photoshop4 and instructions from Rich Media5 to overlay color-coded boxes on the library floor plans. Occupancy percentages were translated into colors using a gradient of blue (low occupancy) to red (high occupancy) based on those used in Khoo3 and Gullikson.2 Results Studying the use of space in a library over two weeks does not give a complete picture of how the space is used throughout the entire academic year. However, using heat maps allowed us to identify some interesting trends. Table 1 illustrates the total use of the library by floor during both the high occupancy week and the low occupancy week. In general, use of the library was slightly more than double during the high occupancy week compared to the low occupancy week. The Second Floor was the most used area of the library, regardless of the time of day. Compared to the low occupancy week, the First and Second Floors were nearly twice as 2 Peer Reviewed Article Hypothesis , Vol. 35, No. 1, 2023 popular as the high occupancy week, and the Lower Level was nearly three times as popular. The heat maps made these differences easily visible. Table 1. Occupancy counts and percentages from two one-week periods High occu- pancy week count Occupancy percentage Low occu- pancy week count Occupancy percentage 2nd Floor 1,854 20.74% 846 9.46% 1st Floor 743 14.04% 417 7.87% Lower Level 1,145 17.04% 457 6.80% Total 3,742 17.85% 1,720 8.21% As seen in the heat map below (figure 1), the Lower Level was quite popular during the high occupancy week, with some surprising findings. The study tables in Zones 6, 7 and 8 were very popular. Our assumption was that this was due to their proximity to the only sources of natural light on this level. However, the carrels in Zone 2 - which are furthest from all natural light sources – were almost as popular, if not more so, during the same time period. Additionally, patrons seemed to prefer more secluded areas of the floor with less foot traffic, such as near the stacks. Figure 1 clearly shows the study rooms in Zone 3 and carrels in Zone 2 were used more often than other areas on the Lower Level. Perhaps the Lower Level is more popular during times of intense studying overall due to its seclusion and the large number of seats located near natural light. Figure 1: Lower Level during high occupancy time period 3 Peer Reviewed Article Hypothesis , Vol. 35, No. 1, 2023 Figure 2, illustrating the data on the Lower Level from the low occupancy week at the beginning of the semester, appears much cooler (more blue areas) than in figure 1. By comparing the two figures, the difference in colors clearly illustrates how the same library space is populated at different times of the year. Of particular interest is that Zone 3 was popular during both times of the year indicating a strong preference for this location. Figure 2: Lower Level during low occupancy period Using heat maps also made it clear that some high-traffic areas were never popular. Zone 11 contains a large circular staircase that runs through all three floors with nearby seating. That seating was not popular in either week, perhaps due to the disruptions from people using the stairs. One of the key benefits of using heat maps is the ability to easily identify which areas of a library are popular or unpopular just by noting how ‘hot’ (bright green/yellow) or ‘cool’ (dark blue) an area is colored. Analyzing the maps and the underlying data also identifies other potential topics for more study and consideration. The key limitation of the heat map method is understanding why a particular space in the library is popular or unpopular with patrons. Given this limitation, the heat maps generated in this study will be used to help inform space use decisions, but only in conjunction with direct feedback from library users gathered via focus groups and advisory groups. 4 Peer Reviewed Article Hypothesis , Vol. 35, No. 1, 2023 This method did not answer all our questions, and often raised more, but provided a visual way to easily understand library space usage across multiple time points. Further Reading Given LM, Leckie GJ. (2003). “Sweeping” the library: Mapping the social activity space of the public library. Library & Information Science Research. 2003;25:365–385. https://doi.org/10.1016/S0740-8188(03)00049-5. Nitecki DA. (2011). Space assessment as a venue for defining the academic library. The Library Quarterly. 2011;81(1):27-59. https://doi.org/10.1086/657446. Schmidt A. (2016). Data-driven design. Library Journal. 2016;141(6):26. http://lj.libraryjournal.com/2016/04/opinion/aaron-schmidt/data-driven-design-the-user- experience. References 1. Khoo M, Rozaklis L, Hall CE, Kusunoki D, Rehrig M. Heat Map Visualizations of Seating Patterns in an Academic Library. IConference 2014 Proceedings, Jan. 2014. https://doi.org/10.9776/14274. 2. Gullikson, S, Meyer, K. Collecting space use data to improve the UX of library space. Weave, Journal of Library User Experience. 2016;1(5). http://dx.doi.org/10.3998/weave.12535642.0001.502. 3. Khoo MJ, Rozaklis L, Hall C, Kusunoki D. (2016). “A really nice spot”: Evaluating place, space and technology in academic libraries. College & Research Libraries. 2016;77(1):51-70. https://doi.org/10.5860/crl.77.1.51. 4. Adobe. Adobe Photoshop User Guide [Internet]. Create data-driven graphics. https://helpx.adobe.com/in/photoshop/using/creating-data-driven-graphics.html [Cited 27 Apr 2021]. 5. Rich Media Creative Services. Using variables in Photoshop with external data source. http://www.richmediacs.com/user manuals/RMCS PS Training/Using%20PS%20Variables /UsingVariablesInPS EXTERNAL.html [Cited 8 Aug 2021]. 5 https://doi.org/10.1016/S0740-8188(03)00049-5 https://doi.org/10.1086/657446 http://lj.libraryjournal.com/2016/04/opinion/aaron-schmidt/data-driven-design-the-user-experience http://lj.libraryjournal.com/2016/04/opinion/aaron-schmidt/data-driven-design-the-user-experience https://doi.org/10.9776/14274 http://dx.doi.org/10.3998/weave.12535642.0001.502 https://doi.org/10.5860/crl.77.1.51 https://helpx.adobe.com/in/photoshop/using/creating-data-driven-graphics.html http://www.richmediacs.com/user_manuals/RMCS_PS_Training/Using%20PS%20Variables/UsingVariablesInPS_EXTERNAL.html http://www.richmediacs.com/user_manuals/RMCS_PS_Training/Using%20PS%20Variables/UsingVariablesInPS_EXTERNAL.html