Chapter 3 — Graph and the Chart-Pencil Set

Graph settled onto her stool, a tiny finch with yellow-and-cream feathers streaked with warm russet. Her bright eyes scanned the classroom, taking in every detail. From her vest pocket, she pulled a small leather case. It held eight colored pencils, each a different shade, nestled in neat rows. Next to them, a small folded card listed chart types: bar, line, scatter, pie, histogram, box-plot, heatmap, map. Graph liked things precise.

When a new set of numbers appeared, she would unfold that card. She chose the right tool for the data’s shape. Then she’d draw the chart, picking a color that matched the data’s mood. This was her art, her science: data visualization. It was more than just making pretty pictures. It meant choosing a chart that showed the truth about the numbers.

Many people, especially beginners, picked charts that looked cool. They chose the flashiest one, or the one that filled the most space. Maybe it used all the colors of the rainbow. But Graph knew that was a mistake. The best chart wasn’t about looking impressive. It was about matching the chart’s form to the data’s true shape.

She held up two small drawings. “Look at these,” she chirped. “If you want to compare how many students prefer apples versus oranges, you use a bar chart.” She pointed to a simple drawing with two bars of different heights. “Bars show categories. They let you compare different groups.”

“But if you want to see how the number of students changed over the school year, you need a line chart.” She showed another drawing, a wiggly line rising and falling. “Lines show change over time.”

“What if you want to see if taller students also tend to have bigger feet?” a student named Pip asked.

Graph smiled. “Excellent question, Pip! For that, you’d use a scatter plot. It shows relationships between two different variables.” She quickly sketched dots on a grid. “And if you want to show how parts make up a whole – like what percentage of our school is fifth graders, sixth graders, and so on – you use a pie chart. But only if those parts add up to one hundred percent.” She drew a circle sliced into wedges. “Each chart has a specific job. And the data always has a shape that matches one of those jobs.”

Graph never thought of visualization as decoration. “The chart IS an argument,” she would say, her voice firm. “Every chart choice claims something about the data. The chart you pick frames how the viewer sees the data. The chart that tells the truth is not always the loudest one.”

This mattered because the popular, flashy charts—the ones with 3D effects, silly animations, or rainbow color palettes—often hid the data. They obscured it rather than revealing it. Graph wanted her students to see charts as truth-serving claims, not just attention-grabbing displays.

She didn’t just teach them how to make good charts. She also showed them how to spot the bad ones. “Some charts lie,” she told her class, her voice serious. “Not on purpose, maybe, but they hide the truth.” She projected an image of a bar chart. The bars looked hugely different in height. “See this? The y-axis, the one that goes up and down, doesn’t start at zero. It’s truncated.” She drew a line on the screen, extending the axis down. “When you start it higher up, small differences look enormous. It exaggerates things.”

She showed another chart, with two lines wiggling across it. “This one has dual y-axes,” she explained. “Two different scales, one on each side. If those scales don’t match up, it can make things look like they’re related when they’re not.” She pointed to a 3D pie chart. “And these 3D distortions? They just make it harder to compare the slices. They look fancy, but they confuse the eye.”

“Sometimes,” she continued, “people only show you a cherry-picked time window – just a small part of the data that makes their point look good. Or they forget to give you a baseline, so you don’t know what normal looks like. The real skill,” she insisted, “is seeing through the chart, back to the actual numbers. Back to the data.”

Graph grew up in a small village nestled among tall, whispering pines. Her family had always been the village’s quilt-makers. Every year, they designed and stitched the annual harvest-quilt. Each square on the quilt showed a household’s contribution to the harvest. The colors weren’t just picked because they were pretty. They were chosen carefully, to honor the data of the village. A deep green might mean a plentiful corn harvest. A bright yellow could show a bumper crop of sunflowers.

Graph remembered watching her grandmother’s nimble claws. Her grandmother would frown at a square where the colors seemed off. “That quilt,” she’d say, “it uses flashy colors arbitrarily. Nobody will trust that quilt.” But the quilts whose colors carried real meaning, those became village heirlooms. They hung in the great hall for generations. By the time Graph was six, she understood. Making a quilt, like making a chart, was a craft of honesty. The quilt that earned trust was the one that truly honored the data.

When Graph was twenty-two, she walked all the way to the DataForge academy. The headmaster, Datum, a wise old owl with spectacles perched on his beak, looked at her intently. “What is data visualization?” Datum asked, his voice a low rumble.

Graph straightened her small shoulders. “It is the shape of the story,” she said. Her voice was clear, even for a finch. “The chart that tells the truth is not always the loudest one. Every chart choice claims something about the data. We must match the chart to the data’s shape. And then, we teach the viewer to see through the chart, back to the data itself.”

Datum blinked slowly. A small smile touched his beak. “You are appointed,” he said.

In her workshop, Graph began every first-day lesson the same way. She unfolded her chart-type-reference card. She opened her chart-pencil case. Then she looked at her students, her bright eyes serious. “I am Graph. The data-pipeline primitive I teach is visualization. The move is match the chart to the shape of the data.”

She pointed to her card. “Bar for categories. Line for time. Scatter for relationships. Pie for parts-of-a-whole. Histogram for distributions.” She paused, letting the words sink in. “Each chart has a job. Pick the chart that fits.”

She taught them the steps for good visualization: First, identify the data’s shape. Is it categorical, like types of fruit? Is it continuous over time, like daily temperatures? Is it geographic, showing locations? Knowing the shape helps you choose. Second, match the chart-type to the shape. If it’s categories, use a bar chart. If it’s continuous over time, use a line chart. If it’s two continuous variables, a scatter plot. Third, resist the flashy default. “Those 3D charts? Usually worse than 2D charts at showing data,” Graph explained. “Rainbow color palettes? Usually worse than purposeful color choices. Pie charts with more than five categories? Usually worse than bar charts.” Her students often chuckled at the thought of a pie chart with twenty tiny slices. Fourth, start the y-axis at zero (usually). “Truncated y-axes exaggerate small differences,” she warned. “Exceptions exist, like temperature data, but they need explicit labeling.” Fifth, label everything. “Axes, units, time-range, data source, the number of items included,” she listed. “The chart should tell the viewer what they’re looking at without you having to explain it.” Sixth, test the chart by asking ‘what does this chart claim?’ “If the claim is not warranted by the data, the chart is misleading,” she said. “Redesign it.” Seventh, teach the misleading-chart patterns explicitly. “So you kids can spot them in adult life,” she said. “Truncated y-axes, mismatched dual y-axes, 3D distortion, cherry-picked time-windows, missing baselines. Learn to see them.” Finally, she reminded them: “The chart IS an argument. Visualization is not neutral display. Every chart choice claims something. Make the claim honest.”

“I sometimes draw a chart that looks beautiful but obscures the data,” Graph admitted to her students. “That’s not failure. That’s how I learn to spot when aesthetic and honesty are pulling apart. The redesign is the practice — honesty first, then beauty.”

When students asked Graph whether visualization was hard, Graph always said the same thing:

“It is not hard. It is match the chart to the shape of the data. The chart that tells the truth is not always the loudest one.”

She closed the chart-pencil case with a soft click. The next dataset waited to be charted.

---