We continue our journey into the world of data visualization. Not long ago we broke down the Sunburst chart, and today it is time for three tools that appear in analytics more often than any others — charts, histograms, and the Pareto chart. At first glance they look similar: they all have bars. But each one answers a different question. A histogram shows distribution, a Pareto chart shows priorities, and a regular chart shows comparison of categories. Once you learn to tell these three tasks apart, you will stop building "bars for the sake of bars".

The data is the same real metrics from the devitpro.uz website in the Cloudflare dashboard. This time the period is 3 October — 1 November 2026. Over 30 days we recorded 39,726 requests (Total requests) and 6,656 unique visitors (Unique visitors). And the most interesting part: on 9 October there was an anomalous spike — 13,163 requests in a single day. It is on data like this that the Pareto rule shows up most clearly.

Key idea: the Pareto chart is not a separate chart type — it is a combination of a histogram and a cumulative curve. It answers the question "which 20% of causes produce 80% of the result?" and turns raw data into a list of priorities.

The Data Behind the Charts

We have two Cloudflare datasets for the period from 3 October to 1 November 2026. The first is Total requests (total number of requests per day). The second is Unique visitors (unique visitors per day). Below are the key figures.

39.73k
Total requests
Total requests for 3 OCT — 1 NOV 2026
6.66k
Unique visitors
Unique visitors over 30 days
13,163
Peak requests
9 October 2026 — anomalous day
344
Peak visitors
18 October 2026 — period maximum

Histogram: What Distribution Really Means

A histogram is a way to see how values are spread across ranges. Unlike a regular bar chart, where the X axis holds categories (countries, devices, browsers), in a histogram the X axis holds value intervals (for example, "0–500 requests", "500–1,000", "1,000–2,000" and so on). The height of each bar shows how many days fall into each interval.

Why is this useful? A histogram instantly answers the questions: "What does a typical day look like?", "How spread out is the data?", "Are there outliers?". If all bars are grouped in one place, the process is stable. If the data is "smeared" across a wide range, there is instability or anomalies.

How a Histogram Is Built

  • Step 1. Collect all values (in our case, 30 daily request values).
  • Step 2. Find the minimum and maximum — these are the range boundaries.
  • Step 3. Split the range into equal intervals (bins). 5–10 bins is optimal for 30 points.
  • Step 4. Count how many values fall into each interval and draw the bars.
Sturges' rule — number of bins k = 1 + 3.322 · log₁₀(n)
where n is the number of observations. For n = 30: k ≈ 6
1. Histogram: distribution of daily requests (without the outlier)
Built from 29 normal days. The anomalous spike on 9 October (13,163 requests) is excluded so the scale is not compressed — it is shown separately.
13 days — 400–700 requests · 11 days — 700–1k · 9 Oct (13,163) — outlier, outside the histogram

The histogram shows the main point: a typical day is 400–1,000 requests, and those two bins hold 24 of the 29 normal days. The distribution has a visible "hump" in the middle and a light right tail (days with 1,000–1,600 requests). And 9 October is a separate story: 13,163 requests, an outlier that is almost 10 times larger than the rest. Outliers like this usually come from either viral content or anomalous activity (bots, scanning, a DDoS attempt). The histogram does not answer "why", but it instantly shows "what" and keeps a single day from ruining the whole picture.

Chart: Comparing Categories

A regular bar chart is the most underrated tool. Its strength lies in comparing categories with one another. The X axis holds categories (days, countries, browsers), the Y axis holds the value. No math — just "which is bigger, which is smaller".

In our case a great task for a chart is comparing the top 10 days by the number of requests. This is not a histogram (there are no intervals here) and not a Pareto chart (there is no accumulation here) — it is pure comparison: which day was most active, which was second, and how big the gap is between them.

2. Chart: top 10 days by number of requests
Comparison of categories without accumulation. 9 October dominates — 13,163 requests, about 8 times the median day.
9 Oct — 13,163 · 17 Oct — 1,583 · 4 Oct — 1,408 · 9 Oct is far ahead

The chart shows that 9 October is not just "first place". It is a gap of about 8 times from second place. A comparison like that immediately suggests an idea: traffic analysis needs two different approaches — one for "normal days", another for "anomalous" ones.

Pareto Chart: The 80 / 20 Rule in Action

And now for the main part. The Pareto chart is a histogram whose bars are sorted in descending order, plus a cumulative percentage line. It is named after Vilfredo Pareto, who noticed that 80% of the land in Italy belonged to 20% of the population. In analytics the rule is usually phrased like this: roughly 20% of causes produce about 80% of the result.

How a Pareto chart is built:

  • Step 1. Take the categories and their values (in our case, days and request counts).
  • Step 2. Sort the categories in descending order by value.
  • Step 3. Draw the bars in that order — this is the histogram.
  • Step 4. Add the cumulative percentage line (the cumulative curve).
  • Step 5. Find the point where the curve crosses 80% — that is the "vital few".
Pareto cumulative percentage Cum%ᵢ = (Σ value₁..ᵢ / Σ value₁..ₙ) · 100%
where i is the position after descending sort, n is the total number of categories
3. Pareto chart: requests by day + cumulative curve
Bars are daily requests, sorted in descending order. The line is the cumulative percentage. The 80% crossing is the key point.
Hover over the line to see how many days produce 80% of traffic.

What Our Pareto Chart Showed

After sorting the days by descending requests and building the cumulative curve, we found that just one day (9 October) accounts for about 33% of all requests. And three days — 9, 17, and 4 October — already account for more than 41% of traffic. In other words, the 80 / 20 rule works here in an even harsher form: roughly 10% of days (3 out of 30) generate 40%+ of traffic.

33%
One day (9 Oct) alone produced a third of traffic
41%
3 out of 30 days produced 41% of requests
80%
Pareto threshold — reached in fewer than 10 days

This is exactly the power of the Pareto chart: it does not just show "9 October had a lot of requests". It quantitatively proves that traffic analysis can be radically simplified — focus on a handful of key days and figure out what happened on 9 October. The other 27 days are "useful but noise".

Important caveat: the Pareto rule is a heuristic, not a law of nature. It does not always produce exactly 80 / 20. In our case we got 40 / 10 — and that is fine. The main point is not to force the data to fit the rule, but to use the rule as a prioritization tool.

And What About Visitors? A Second Look Through Pareto

It is interesting to look at Unique visitors separately. The picture here is different: the peak is on 18 October with 344 visitors, while 9 October, the "king of requests", showed only 249 visitors. This is a classic example of why requests and people must never be confused.

On 9 October, with 13,163 requests, there were only 249 unique visitors — that is about 53 requests per person. It looks a lot like automated activity: a bot, scanner, or parser. And on 18 October, 344 visitors generated 694 requests — that is 2 requests per person. Normal human activity.

4. Pareto chart: unique visitors by day
A very different picture: the distribution here is much more even, with no giant outliers.
18 Oct — 344 · 8 Oct — 283 · 4 Oct — 275. Spread is 2.3×, not 37×.

Compare: for requests, the spread between maximum and minimum is 37 times (13,163 vs. 358). For visitors it is only 2.3 times (344 vs. 151). That is the key takeaway: people behave steadily, bots do not. The Pareto chart for visitors shows a much flatter curve — the 80 / 20 rule is weaker here.

What I Took Away From This Experience

  • Histogram — for distribution. It answers "how are values spread out?", not "which category is bigger?".
  • Chart — for comparison. Simple, honest, no accumulation. Perfect for top 10 lists.
  • Pareto — for prioritization. It combines bars and a cumulative curve and shows where "the important" ends and "the rest" begins.
  • Never confuse requests and people. Anomalies in requests are often caused by bots. Always check the second metric.
  • The Pareto rule is a heuristic. It helps you think, but it does not replace analysis. 40 / 10 is a result too.
  • For SEO, text matters. Search engine bots do not read canvas charts. All key figures and conclusions must live in the text — which is what we are doing.

Verdict: Is It Worth Learning These Three Tools in 2026?

The answer is a resounding yes. The histogram, the chart, and the Pareto chart are the basic vocabulary of any analyst. Without them you cannot describe data, find anomalies, or set priorities. And the most valuable part: they work not only in web analytics, but also in finance, manufacturing, marketing, and machine learning — wherever there is data and there are decisions to make.

For me personally, this breakdown was an important lesson: the same dataset can be shown three different ways, and each one reveals something different. The histogram showed the outlier, the chart showed the gap, the Pareto chart showed the concentration. Visualization is not "decoration" — it is a way of thinking.