Viz ideas to L.A.T.C.H. on to
L.A.T.C.H. is an approach for helping you organize any dataset and is a fundamental part of information design. So let's say you’ve happened upon a rich dataset, and you’re sure it has great potential. But you’re left staring at it wondering where to even begin. Or perhaps you feel like you're in a visualization rut: you're doing the same thing over and over. As presented at the 2020 North American Cartographic Information Society (NACIS), follow me through some viz ideas for your dataset that you can L.A.T.C.H. on to!
You can thank Richard Saul Wurman for the solution to your problem. Perhaps you've heard the name before? He is the founder of TED conferences and has written, designed, and published more than 90 books, including Information Anxiety, released in 1989 and revised in 2000. Information Anxiety is the source for answers you seek to your visualization needs.
You Haven’t Told Me Really Anything Yet That Will Solve My Problem!
Well, I can tell someone’s feeling a bit anxious! Take a deep breath and find your calm for a minute. The answer is in what Wurman calls “The Five Hat Racks” or the five different ways information can be organized and visually presented. These five methods for arranging information can be turned into the easy-to-remember mnemonic L.A.T.C.H.
L is for Location
Data can be organized by location—where it exists in space. This most often lends itself to a map or other spatial diagram.
A is for Alphabet
Also consider organizing information alphabetically. It’s as easy as ABC. Usually a table or matrix of some sort.
T is for Time
Think of a timeline: how does your data unfold chronologically?
C is for Category
Can your data be grouped by common characteristics? This would organize your data by category. Visually this often comes in the form of a chart or table.
H is for Hierarchy
With what magnitude do your data points show up? Less vs. more? Better vs. worse? Some sort of ranking makes hierarchy work. Wurman also calls this “Continuum,” but that would make the acronym an unwieldy L.A.T.C.C., so let’s stick with hierarchy and L.A.T.C.H.
Examining two datasets should illustrate well how L.A.T.C.H. works. The first dataset is my children's annual Halloween candy haul, and the second is all NACIS annual conferences held since its founding in 1980.
By late evening on each October 31st, my small brood of six children has filled the home with pounds and pounds of candy. It’s all dumped on the ground and evaluated with amazing rapidity. Then the unwrapping and consuming commences. My wife and I are able to eventually stop the sugar abuse to the pancreas and liver and get the children to bed. But what happens the next day? Another candy binge? No! It’s a Halloween candy L.A.T.C.H. party! Each year the children come up with different ways of organizing their candy haul into data visualizations. (After the L.A.T.C.H. party—though this has nothing to do with L.A.T.C.H.—the Halloween fairy takes away all the candy and leaves some other not-so-sugary fun behind.)
Last year, the candy was mapped by where it had been invented. We used our awesome map of the United States—hung in the dining room to encourage dinnertime geography conversation ;)—to tape each candy wrapper in its proper location. One year the children mapped each house in our neighborhood, coloring whether it handed out candy at the door, on the porch, or not at all.
Well, this one’s pretty straightforward. Place the candy in alphabetical order from Dum-Dum to Tootsie Roll. The more candy, the greater the fun.
This data collection has ranged from the year the candy was invented to how long it took to eat the candy. I don't have a photo timing ourselves as we chow down on chocolate, so a still from the old Tootsie Pop commercials will have to serve as a stand-in: how many licks does it take to get to the center of a Tootsie Pop?
This may be the data organization strategy with the most fun. How about this formula?
Masking tape + Venn diagrams = an amazing array of categorical fun
You probably can't read the categories written on the tape, but the three inner circles are: nougat, nuts, and caramel. The outer circle is chocolate. A representative wrapper from each kind of candy is then placed appropriately in the Venn diagram. Snickers gets the honored place in the middle with chocolate, nougat, nuts, and caramel. One of my favorite candy bars, PayDay, makes the Venn diagram extra tricky, because you need the intersecting circles of nougat, nuts, and caramel to also be outside the chocolate ring. Apparently no one got PayDays in this haul. :(
The Venn diagram is usually slightly different each year. Another year, we wrote out on paper what the different categories were. This has all the same categories as the previous, but adds crispies and candy shell categories as well.
The main hierarchy activity each year begins with a categorical activity: sort and chart your candy by color wrapper. You can see on this chart that red, brown, and silver are big candy bar colors. Next, each wrapper color is randomly assigned a point value. Then, each child calculates the points for each color category and their cumulative score. One child is crowned the (very arbitrary) winner!
One year, the children expressed grave concern that their favorite Skittle flavor—strawberry—occurred with less frequency than the other flavors. So they embarked on another hierarchy exploration. All 23 bags of Skittles were opened, categorized by flavor, and averaged by how frequently they showed up in each bag. The result? Out of 23 bags, strawberry had the lowest average frequency with 2.6 per bag. Lemon had the highest with 3.1 per bag. We didn't do any advanced statistics to determine statistical significance—the children were too young then, though some of the older children might really enjoy that this year! We just averaged everything out to 3 per color per bag and called it a day. :D
Candy Is Sweet and All, But What About Mapping Data?
You’re right! Let’s take the concept to a dataset that might be more along map lines. Since 2020 is the 40th anniversary of NACIS, how about we use a dataset of all NACIS annual meetings? I compiled this dataset and invite you to use it and explore it as you like!
A Summary of the Dataset
This was a really fun dataset to compile and explore. The real excitement of putting your data through a L.A.T.C.H. exercise is when you start to see patterns you might not have noticed if you had just stuck with your standard approach. You also see how some of the organizational strategies might even blend in with each other—for example, how visualizing NACIS conferences by decade might be a Time and Category approach.
Here's a high-level review of the dataset, followed by a list of the fields I created.
41 annual meetings (including 2020)
34 different host cities
39 different venues (only two venues have been repeated—Hyatt Regency Milwaukee and Portland Downtown Doubletree)
2 sites have been converted into senior living centers (Ann Arbor and St. Paul)
2 sites have been torn down and replaced by hospitals (Philadelphia and Atlanta)
1 site was not a hotel (University of Ottawa)
No. of Days
No. of Presentations
W of Mississippi
I also really wanted to include attendance and generated income as great hierarchical fields, but, alas, that data was too difficult to come by.
The Maps and Other Data Visualizations From the Dataset Using L.A.T.C.H. Principles
Some of the real excitement in data visualization is when you start to combine different approaches to looking at data. The following infographics mix and match L.A.T.C.H. in different ways. Click on the images to see them bigger!
L.A.T.C.H. is perfect for exploring a brand-new dataset or to help you get out of a data visualization rut. I've tried to keep most of the visualizations in the map realm, since this is for a cartography conference, but be courageous and explore the gamut of different visualizations—charts, diagrams, timelines, tables, matrices, and on!