Project Phases and Planning

This research consisted of seven phases: data collection planning, first-round data collection, second-round data collection, data synthesis, personas development, reflection, and presentation and integration. Here is a brief description of each phase.

Phase 1: Data collection planning

• Decide on primary method of data collection as a semi-structured interview.

• Curate a list of users to request for interviews.

• Contact users identified to interview in March and April 2019.

• Draft guiding interview questions.

Phase 2: First-round data collection

• April 2019: Conduct nine semi-structured interviews that yield usable data (primary data).

Phase 3: Second-round data collection

• Identify geospatial reference interactions with faculty between January 2018 and May 2019.

• Use the information from the reference data to supplement interview data (secondary data).

Phase 4: Data synthesis

• Transcribe interview notes and structure them into a spreadsheet.

• Identify themes and concepts within the primary and secondary data using online tools for text and content analysis.

Phase 5: Personas development

• Design personas round 1: Draft and storyboard text-based personas.

• Design personas round 2: Design visual personas.

Phase 6: Reflection

• Compare personas with data.

• Refine personas as needed to better reflect typical user-interactions based on the data.

Phase 7: Presentation and integration

• Create an internal report to present to the library team and administration.

• Integrate improvements identified from personas into annual planning, service improvements, programming, and budget justifications.

Data Collection

We collected the primary data used in the development of the personas through semi-structured interviews and supplemented with secondary data from recorded geospatial reference consultations with the target group from the previous year and a half. We used this time range for the secondary data because no other experiential data was available; the geospatial data services librarian entered the role in January 2018 and no well-documented or recorded reference information related to geospatial consultations was available prior to that.

The semi-structured interviews were with doctorate-holding faculty, researchers, postdocs, and instructors at the university. We chose this method because library colleagues had shared that they had seen low response rates and minimal success with more impersonal surveys and had better success with in-person contact. Additionally, interviews allowed the people to share more personal and in-depth details about how they integrated geospatial data and tools into their research, which produced a significant set of qualitative data we could use to enhance the program’s services.

We contacted the users for interviews based on previous geospatial reference data and recommendations from colleagues with departmental liaison roles. The initial outreach requests for interviews targeted roughly representative responses from each academic division, which included arts and humanities, social sciences, science, and engineering. A limitation of our study is that the interview request response rate and interest across academic divisions varied, which is an expected challenge when collecting data from people whose busy schedules may not be able to accommodate time for interviews.

We developed interview questions to guide the conversation, but due to the semi-structured nature of the interviews, not all interviewees answered every question. The prepared interview questions were:

• Would you consider yourself a current user of GIS or geospatial data?

• If yes, how do you use it?

• What have your prior institutions had that we do not?

• What would you like to see?

• Are there any geospatial needs you have that are currently not being met that you think the university or library could be meeting?

• Do you think GIS/geospatial data are important to your colleagues, deans, administrators, students?

Exploratory Data Visualization

We used the online environment Voyant Tools (https://voyant-tools.org) to generate exploratory data visualizations from both the text of interviews and supplemental data. This led us to better understand the data and derive general categories and themes to use across all the personas. Voyant Tools is an easy-to-use application that provides content and text analysis to generate word counts and pairings, as well as useful visualizations, that demonstrate common themes and connections between words. Major themes from the interviews included various ways users could receive geospatial data and research support, student training, funding, and labs.

For example, Figure 1 is an image generated from the Collocates Graph tool in Voyant Tools. It shows the selected keywords GIS, data, and research in blue and words that are often collocated with, or nearby them, in orange. This visualization helped generate themes and categories for realistic researcher profiles, goals, and frustrations for the personas.

Figure 1

Theme visualization.

One way to interpret the link based on the keyword “GIS” is that some interviewees who were using GIS were focused on publishing papers and disseminating results. From the “research” keyword, we could interpret that some people needed continued research support not only for themselves but also for capstone projects, which all seniors are required to complete. The social link could reflect a theme many interviewees touched upon: the desire for a community of practice at the university surrounding geospatial research. Lastly, the keyword “data” and its links might best reflect themes from both primary and supplemental data where people needed assistance with data collection, turning data into maps, or learning more data skills to complete geospatial research projects. When interpreting this visualization, our understanding of both the context of the interviews and supplemental data is essential to make sense of the resulting image.

Determining How Many Personas to Create

There is not a specific desirable number of personas to create. In this study, we identified three groupings of characteristics in the primary and secondary data. Broadly, the people interviewed or interacted with in the supplemental data could be described as people who did not see themselves as GIS users but wanted to use it more, those who did not see themselves as GIS users but generated significant amounts of geospatial data in their research, and those that were frequent users of geographic information and GIS and would identify themselves as such. These groupings led to the creation of three distinct, realistic, and well-rounded personas that capture a wide cross-section of users as represented by the data available. Certainly, there could be more groupings, but from the data available there was too little data to create more than three holistic personas. Additionally, three were also enough to provide a consistent and actionable story without overwhelming stakeholders when we reported the study.

Persona creation is basically storytelling, so while it might be necessary to fill in gaps when creating personas, it defeats the purpose to concoct a story that is no longer rooted in the data just to create more but less accurate or false personas. For instance, due to the minimal representation from the engineering division in the data, you will see in the following sections that there is no persona with a role in engineering despite the existence of this user population; no persona is better than an inaccurate persona.

Persona Design Round 1: Storyboarding the Personas

After reviewing the data, we storyboarded the personas using six elements for each. We chose these elements based on features typically used in personas (Harley, 2015). Some basic building blocks of personas are name and tagline, a short biography, experience with the service or tool, goals and motivations, frustrations, and areas of intervention or service improvement. In this case, we centered the elements around GIS usage and in a way that other library stakeholders, who may not be knowledgeable of GIS, would be able to understand.

The elements of these personas are:

1. Name and tagline.

2. GIS research profile: a biography to describe their role at the institution and research background.

3. GIS technical ability.

4. Motivations and goals for using GIS data.

5. Frustrations or challenges to using GIS data and tools.

6. Helpful library interventions.

To create realistic profiles, we gave all personas a name, a faculty role within an academic division, and a background of previous knowledge or experience with GIS data and tools. While not all interviewees were faculty, all personas had at least a pseudo-faculty role, such as lecturer, for a clearer presentation because the research of faculty and non-faculty members often overlaps. In the future, one could add a postdoc or research associate persona to this series. We also used the personas to note the ways the library was already positioned to support high-level and long-term research as well as areas of potential growth and improvement.

We did not use visuals for the first round of persona design; we thought of them as text-based outlines to eventually aid in the visual design process. The initial personas are represented in Appendix B.

The three personas are Rachel the Novice, A.J. the Professor, and Jean-Paul the Big-Thinker. We chose the names and characteristics based on demographics and traits of the interviewees, what was represented in the supplemental data, and our university’s academic population. An actual user of geospatial services at the university might not fall into every category of a single persona, and there is no reason for users to directly mimic a persona. For example, a full professor asking for first-time assistance with geospatial data likely has more shared characteristics with Rachel the Novice than Jean-Paul the Big-Thinker, even though Jean-Paul is a full professor as well. Put differently, one should not assume all full professors have the same geospatial research needs as The Big-Thinker even though they share an equivalent academic rank.

Persona Design Round 2: Visual Design of the Personas

The purpose of the visual output of the personas was to convey user-information in an easily understandable format to non-GIS services providers and stakeholders in the library. The information needed to be persuasive and believable to library administrators. Most importantly, it needed to be a reference that we could return to when thinking about how to design programming and support research. An additional objective of creating personas was to provide any future librarian within the GIS program with a reliable foundation to serve the university’s diverse community of users.

We used a free, personal Canva (https://www.canva.com) account to design the personas visually. Canva is a freemium, web-based tool for design. We formatted each of the six elements within a resume template and touched them up with features such as free head-shot style photographs.

We chose the photographs to provide a sense of realism to the personas. With a free Canva account, limited head-shot style photographs were available. Nevertheless, it was important to select images that reflected the diversity of our user population. We did not settle for images of only one demographic, as it would not be reflective of all or most faculty and researchers at the university. We tried to be as inclusive as possible with the limited means available, and with each image, we chose one to reflect personality traits one might discern from looking at the image. While some researchers choose not to include images in their personas (Miklaszewicz, 2019), we chose to include them because it would be important for making the personas memorable to stakeholders.

Jean-Paul the Big-Thinker

Jean-Paul the Big-Thinker is represented as an older man in a relaxed pose. The person in the image has a far-off, visionary look in the eyes to convey the long-shot ideas and big dreams heard from some of the interviewees.

Beyond the visual design, there is another difference between the first and second rounds of the personas. In round one, the geospatial technical abilities were more nuanced for each persona, but we condensed these in the final visual report. For example, in the second round, we list GIS technical skills but show a scale of ability with ArcGIS. While ArcGIS may not be the primary GIS tool for everyone, we chose it due to its brand-name familiarity among the library stakeholders. The second round, essentially, summarized the nuanced text of the first round into actionable information for colleagues and administrators to easily digest. Figures 2–4 show the final products.

Figure 2

Rachel the Novice

Figure 3

A.J. the Professor

Figure 4

Jean-Paul the Big Thinker

Presenting Personas and Integrating into Work

It’s important that persona information be disseminated to colleagues, supported by administration, and implemented in programming. To this end, we wrote an internal report that outlined the project timeline, methods, discussion of themes from the interviews, and further work to integrate the findings. At a regular meeting of librarians and department heads, the GIS librarian presented the findings to colleagues. The GIS librarian was both creator and consumer of the personas, and the feedback and ideas in the personas were integrated into annual planning. For example, we set goals of creating a new faculty geospatial data consulting program and starting a geospatial community of practice.

Operational and UX-Related Aspects of Specialized GIS Services

Personas must serve a purpose that can answer a question: what interventions can the library offer to assist users in meeting their information needs? This study makes the case that narrow-scope personas are an effective way to visualize typical users for specialized research services, such as GIS data services. These personas conceptualize some typical interactions between a faculty member or researcher and the GIS librarian. If we can relate a faculty or doctorate-holding user to one of three personas, it is easier for us to anticipate an appropriate level of service. For instance, Jean-Paul the Big-Thinker does not need the librarian to show them basic datasets or resources because they have likely created their own datasets. On the other hand, the librarian does need to show Rachel the Novice where to find basic datasets, while A.J. the Professor might already have the dataset in mind but may need advice on adopting the best coordinate reference system for their next analysis.

The “helpful library interventions” section of the personas provided a framework for designing new services and programmatic goals. To make these interventions actionable, the GIS librarian worked with an administrative supervisor to incorporate some of them as annual goals in the library’s strategic planning. At the time of this writing, a more structured faculty consulting service is in development. Moreover, the need for more staff or student workers to assist with GIS research is now on the record in the internal report; the personas and the data we used to create them can serve as justification for additional hiring requests.

In addition to looking forward, one can use personas to validate existing successful services. One such example is the continuation of introductory GIS-skills workshops, which are tailored to meet the needs of The Novice persona. We had considered ending these workshops to support more advanced needs, but it is clearer now that they are still needed but may benefit from better marketing toward those who encompass aspects of Rachel the Novice.

Limitations and Further Research

The aim of this study was to create narrow-scope personas based on the ongoing and high-level research needs of a small research-oriented liberal arts university. While the number of interviews was adequate and in line with UX methods, a significant limitation of this study is the lack of representation from the engineering department. This might simply indicate that engineering users are not as in need of the library’s GIS services. That assumption could be problematic without data, and in the future there may be efforts to reach out again to users in the engineering division. We can update internal documentation if new data are acquired later, or we can create an additional persona specific to engineering.

Another limitation is that we studied only a subset, albeit an active and significant one, of the entire GIS research community on campus. It was logical to begin this study with users who set the GIS research agenda and plan courses, i.e., the faculty. Subsequent research could document the GIS data needs of other categories of users, such as undergraduate and graduate students, external communities, administrative users, and many more narrow user populations that the library may seek to serve.

A final limitation to acknowledge is the question of whether personas are a useful methodology at all. Without careful implementation of the methodology, it is possible for poorly developed personas to throw one off track. In this regard, Adkins et al. (2019) conducted a focus group near the final phase of their research. This is common practice to make sure personas are grounded in the reality of data collected rather than assumptions (Pruitt and Grudin, 2003). We cross-referenced the final personas with the primary interview and secondary reference data to ensure authenticity, but there could be a fully iterative approach that included additional formal focus groups, interviews, or decisive testing to reverify or modify the existing personas.

We believe the personas developed in this study currently reflect the typical faculty or high-level researchers at the university who are utilizing geospatial tools and methods. However, persona creation is an ongoing, iterative process of reevaluation to improve or extend specialized library services. In the future, we could collect more data and revisit the personas to ensure accurate representation of the user community.