Happy holidays, data detectives! At Library Research Service (LRS), we’re wrapping up our posts this year by charting libraries that took an extra step for their community in 2022 by providing services that help fight food insecurity. For the first time this past year we asked libraries on the Public Library Survey whether they distributed meals to children or provided one-on-one help for applying to Supplemental Nutrition Assistance Program (SNAP) benefits in 2022. Of the 112 public library systems in Colorado, 31% (35 libraries) reported that they provided at least one of these services to community members, and six of these libraries reported providing both services. A variety of library types from across Colorado reported providing these services, so this holiday season we created a map to show which libraries provided this type of community support in 2022.
Maps are a broadly loved and appreciated way to display data, and there are many different ways to map data (too many to discuss in one post). To begin, this post will cover some basics of mapping data and introduce two common types of maps. In future posts, we’ll explore more types of data maps and advanced mapping techniques.
Deciding on a Map
If the data you’re working with has a geographical component and there is benefit to seeing the locations of the data points and/or how the data relate to each other geographically, mapping the data is a logical visualization method to choose. However, choosing to map the data is just the first of many choices to make. The type of map to use and how to represent the data will depend on numerous factors such as the number of data points, the locations the data relates to (latitude and longitude coordinates, city, county, or region etc.), the distribution of the data points, and how many dimensions of data are being mapped.
Two of the most common and simple map types are symbol maps and choropleth maps. We’ll start with symbol maps and return to choropleth maps later in this post. Symbol maps will consist of symbols placed on specific locations around the map that depict data through their shape, color, and/or size. You might choose a symbol map if your data points do not relate to established geographical borders and do not create illegible, dense clusters on the map. The data points of 35 libraries that provided help applying to SNAP benefits and meals to children in 2022 fit this criteria, so I chose to chart them with a symbol map.
After deciding to use a symbol map, there are still more decisions to make. To decide whether to vary the symbols by size, shape, or color, you’ll again want to consider the number of data points, the density and spread of the data points on the map, and how many dimensions of data are being mapped. Symbol size may be a logical way to depict data if the purpose of the map is to show a greater quantity or number of occurrences in certain locations. The shape of the symbol can be an engaging way to depict data if it falls into categories that relate to easily recognizable, simple icons. Varying symbol shapes can also be used along with color or size to show more than one dimension of data in one map. I chose color to represent the data in Figure A (below) because this data focuses on the type of service provided, not the quantity of services, and different shaped symbols could unnecessarily complicate the map.
Context is Key!
Regardless of how the map depicts the data, it will probably need a legend to show the reader what the symbols’ sizes, shapes and/or colors are depicting. Legends are particularly important for deciphering maps, so be sure to place it in an easy-to-spot location with easy-to-read text.
A legend likely isn’t the only thing a map needs to provide adequate context for the reader. In the words of Joseph Lalonde (Manager, Data & Analytics, Toronto Public Library) at the 2023 Research Institute of Public Libraries (RIPL) conference, “always wrap your map!” Wrapping your map means including, not only a legend, but also a title, source, scale, and labels if applicable. The purpose of these components is to proactively answer many questions the reader may have when looking at a map–questions such as: Where did this data come from? What is the distance from here to there? What do these sizes/shapes/colors mean? And what is the purpose of mapping this data? Providing enough context without overcomplicating a map is a balancing act. As the maps below demonstrate, creating an interactive map where sections of information appear as the reader mouses over certain areas can be a solution to providing adequate context and labels without crowding the visualization.
Figure A shows the location of the main branch for each library or library system that reported providing help applying for SNAP benefits and/or meals to children in 2022. SNAP is a program that provides a monthly benefit for food to low-income households in Colorado. LRS collected this data in part to assist Hunger Free Colorado, an organization working to end hunger in Colorado.
This was the first year this data was collected (the data can be found at LRS.org), so no baseline data has been previously set. Continuing to collect this data in the future could show whether the number of libraries choosing to provide these services is increasing or decreasing which could help indicate the level of need for these services across Colorado. However, this year, libraries were not asked how many people they helped apply for SNAP or how many meals were distributed to children. Meals may have been distributed over the summer when children are not able to access school lunches and snacks or on Fridays in areas moving to a four-day school week, but LRS did not specifically ask when meals were distributed.
Scaling for Success
In Figure A, each dot represents a library and the color of the dot and the legend tell the reader which service/services that library provides. Scrolling over each dot will bring up a tooltip that shares the name of the library, the service provided, and the median household income of the county in which the library’s main branch is located.
Dots are a great default symbol to use because they are simple and universally understood to indicate a location or data point of interest. The size of the symbols also has a large impact on the map. If this map had much smaller dots they would not stand out, and the map would feel empty. If much larger dots were used the locations of the libraries would not be specific enough. As is, the dots in Figure A are not very specific and cover a much larger area than where the library is actually located, but some location accuracy needed to be sacrificed in order for the libraries to stand out on a map of the entire state.
There are also a few places in Figure A where the dots overlap. A bit of overlap between symbols is generally alright as long as each symbol is still clearly visible. For example, if more libraries in the Denver metropolitan area provided these services, the size of the dots would likely need to be scaled down or an entirely different map layout might be chosen to accommodate the dense cluster of libraries. Overlapping symbols can quickly become a problem when using symbol size to represent data. A larger symbol can cover up the data points around it, obscure the location of the actual data point, or cause readers to confuse the symbol’s size with the size of the location. This is another reason color, not size, was chosen to represent the data in Figure A.
One thing missing from Figure A is a scale for the map itself. There is no measure that shows how many miles a certain distance on this map represents. This could be included, but I chose to instead place certain landmarks on the map to help readers orient themselves. In Colorado, distance as the crow flies can be quite irrelevant if the road to get from one point to another travels around a mountain range. Instead of a scale, this map includes county borders, faint topographic features, and major interstates in Colorado. These features and landmarks help the reader orient themselves and understand where a library is located in relation to familiar places. If the audience for this map was not familiar with the size of Colorado, a scale would be helpful.
A Balanced Map
A recurring theme throughout this post is balance. So far we’ve discussed balancing context and information with overcrowding and balancing the size of symbols with location accuracy and the density of data points. These are some of the most challenging aspects of mapping data, but understanding different map types and design options will help you create a balanced map. Maps can lose an audience’s interest if they are either too complex or too simple. It’s hard to convey a story with just a couple of data points plotted on a map. Figure A is fairly simple, so in Figure B (above) I overlaid a choropleth map to visualize median household income for each county in addition to the symbol map.
A choropleth map uses colors or shades over established geographic areas to depict data for those entire areas. Median household income from the United States Census Bureau is measured by county, so in Figure B each county is shaded to represent its median household income. This allows readers to see if services that fight food insecurity are being provided by libraries in areas with a lower or higher median household income.
Figure B was created by following instructions to make a dual-axis map in Tableau. Unlike Figure A, which only shows median household income in the tooltip for each library, Figure B visualizes how the median household income for each county relates to the location of libraries providing the community support services discussed previously. To show these two data dimensions in one map Figure B has two legends–one for the color of the dots and one for the shading of the counties showing median household income. From Figure B it becomes clear that libraries providing help with SNAP applications and distributing meals to children are located in and near counties with both high and low median household incomes.
There are a couple important things to note when comparing these two data sets. First, the area that a library serves does not always correspond with the borders of the county it is located in, so people living in nearby counties could be served by a library regardless of whether its main branch is located in a different county. Secondly, cost of living across Colorado must be considered along with median household income. Even if a county has a high median household income, it’s possible that the cost of living is also high in that area and there will still be residents struggling to make ends meet.
Both Figure A and Figure B have their strengths and weaknesses. The tooltip in Figure A does not allow the reader to quickly understand if median household income correlates with libraries providing these two services. On the other hand, Figure B takes longer to comprehend with two legends and still does not show a clear correlation between median household income and library locations providing these services. Which map to use depends on the target audience, their level of data literacy, and the ultimate goal of the visualization. In many cases, showing multiple dimensions of data in one map is unnecessary and confusing, but sometimes a dual-axis map can communicate intriguing trends.
Strengths to Mapping
Mapping data is not an efficient way to compare exact values, but if your data is geographically based, it can reveal important patterns and trends. Maps can be used for planning, decision making, communicating with external partners, telling a story, and engaging an audience.
Maps may be such an engaging way to display data because they can feel personal to each person looking at them. Similar to how we like to see ourselves represented in art, books, movies, and music, we like to find our places (our home, places we’ve visited, or places we want to visit) on a map. Presenting data on a map allows the viewer to relate the data points geographically with familiar places, which grow engagement with the visualization.
Zooming out to this bird’s eye view of Colorado libraries that are helping community members put food on the table reveals that public libraries across Colorado, both large and small, rural and urban, have helped fight hunger in their community. Stay tuned for posts exploring more ways to build maps with different Public Library Annual Report (PLAR) data next year!
LRS’s Colorado Public Library Data Users Group (DUG) mailing list provides instructions on data analysis and visualization, LRS news, and PLAR updates. To receive posts via email, please complete this form.