Charts: Radar, Radar with Markers, Filled Radar — Learning New Things in IT and AI with Real devitpro.uz Metrics
We keep exploring the world of data visualization. Last time we worked through surface charts. Today we take the next step: radar charts. This is a family of visualizations that answers a very specific question: “what does an object look like across several dimensions at once?” It lets you compare profiles — not isolated numbers, but the shape of the data. We will explore four variants: the basic Radar, the Radar with Markers, the Filled Radar, and a combined view, all built on real metrics from devitpro.uz.
The data is not abstract. It comes from Cloudflare Requests for the period from September 1 to October 1, 2026, and from Vercel Edge Requests for the period from June 28 to September 1, 2026. A radar chart here works not as a “pretty picture” but as an instrument: it shows how balanced the metrics are, where the profile stretches out, and where it collapses. That is exactly how visualization turns a set of numbers into a story about load, audience and infrastructure.
Key idea: a radar chart is not about “having many axes.” It is about profile and shape. One look at the radar says more than a ten-row table: where the system is strong, where it is weak, and how balanced it is overall.
What Is a Radar Chart?
A radar chart (also called a spider chart or web chart) is a way to show multidimensional data as a closed polygon. Each axis radiates from the center and corresponds to one dimension. A point on an axis is the value of that indicator. Connecting the points gives you the profile of the object.
The main difference from a bar or line chart: here what matters is not the absolute value of each bar, but the relationship between the axes. A radar chart answers questions like: “How balanced are requests across hosts?”, “Which region has a smooth profile, and which is lopsided?”, “Do the profiles of two periods match?”
The four variants we will look at differ in level of detail and emphasis:
- Radar — the baseline outline. Shows the shape of the profile without extra detail.
- Radar with Markers — each axis gets marker points. Helps read exact values and compare several profiles.
- Filled Radar — the area inside the outline is filled with color. Emphasizes “volume” and overlap between profiles.
- Combined — outline + markers + fill all at once. Maximum information density for comparison.
1. Radar Chart — Baseline Profile
Let us start with the foundation. The baseline radar chart shows the profile of a single set of metrics. In our case — the distribution of Cloudflare requests across six key countries: the United States, Uzbekistan, the Netherlands, Italy, Germany, and the United Kingdom. Each axis is a country, and the value is the number of requests.
Notice the shape: the profile is heavily stretched toward the United States. This means traffic is concentrated rather than evenly distributed. For infrastructure this is a signal: if the US segment drops, the overall picture will change much more than if any other segment drops.
2. Radar with Markers — Accuracy and Comparison
Markers on the axes are not decoration. They solve a concrete problem: the outline shows the shape, but it does not let you read values precisely. When there are points on the axes, you can visually align a value with the grid and compare two profiles without peeking at the tooltip.
On this chart I overlaid two profiles: Cloudflare Requests and Vercel Edge Requests, normalized across six shared categories. The markers make it clear where the profiles match and where they diverge. For example, in “Asia” the Vercel profile contributes a much larger relative share, while in “Europe” Cloudflare looks more even.
3. Filled Radar — Volume and Overlap
Filling the area (areaStyle) changes perception: instead of a thin line you see the “body” of the profile. This is useful when you need to compare several objects and show how much they overlap.
Here I show two periods of Vercel Edge traffic: Jun–Jul and Aug–Sep. The filled areas overlap, and it is immediately obvious that the August profile is “wider” on most axes. This is not just growth in the total number — it is growth that affected almost every region, not just one or two.
4. Combined Radar — Everything Together
The fourth variant is the most informative. The outline defines the shape, the markers give reference points, and the fill shows volume. Together they let you read the chart on three levels: what the profile is, exactly where the values sit, and how “heavy” the profile is.
This chart shows four Cloudflare metrics in normalized form: Total Requests, Unique Visitors, Uncached Requests, and Maximum Daily Visitors. The profile is uneven: a peak on Total Requests and a clear dip on Unique Visitors. The fill makes that imbalance obvious.
What I Took Away From This Experience
Radar charts are not “charts for beauty.” They are a tool for working with profiles. Here are a few conclusions I drew while building these four visuals.
- Radar — for shape. When you need to quickly see whether a profile is balanced or lopsided, the baseline radar chart works best.
- Markers — for accuracy. If you are going to compare several profiles or read values “by eye,” markers are a must.
- Fill — for volume and overlap. When there are two or more profiles, areaStyle shows the overlap and the “mass” of each one.
- Combination — for storytelling. When you need to show shape, points and volume in a single visualization.
- Limitation: no more than 8–10 axes. A radar chart quickly becomes unreadable. If you have more dimensions, split the chart or choose a different type.
- Normalization is critical. If the axes use different units (requests and percentages, thousands and units), the radar chart will lie. Always bring values to a comparable scale.
One more thing: a radar chart is a bridge between a table and a map. It does not show “where,” but “how.” That is its strength: you look not at individual numbers, but at the pattern behind them.