How to make a histogram online
Paste numbers into the data box or upload a CSV/TXT file. HistogramMaker.app reads commas, spaces, tabs, semicolons, and line breaks, so data copied from a spreadsheet, lab notebook, survey export, or server log usually works without cleanup. If a CSV has more than one numeric column, pick the column you want to plot.
Choose a binning method, then check the preview before exporting. Freedman-Diaconis is the default because it handles skew and outliers better than some simpler rules, but Scott, Sturges, Rice, Doane, square-root, custom bin count, and custom bin width are available when the chart needs a specific shape or interval.
Add the details a reader will need: a title that names the variable, units on the x-axis when they matter, and a caption for the data source or method. Then export the chart as PNG, SVG, PDF, CSV, or JSON.
If the chart looks noisy, try fewer bins or a rule that gives wider intervals. If the chart hides a cluster or a gap you know matters, try more bins or a custom width. The goal is not to make the smoothest chart. The goal is to show the distribution without inventing patterns.
What is a histogram?
A histogram shows the distribution of numeric data. It groups values into adjacent intervals called bins and draws one bar for each interval. The bar height can show count, relative frequency, density, cumulative count, or cumulative percentage.
Use a histogram when you need to see shape, spread, center, skew, and unusual values quickly. Lab results, exam scores, patient ages, response times, income, sales, manufacturing measurements, and survey results are all good candidates.
Make the histogram early in the analysis. A long right tail, a gap between groups, a pile-up near a target value, or a handful of extreme observations can change how you clean the data, choose a model, or explain the result.
A histogram is not a replacement for summary statistics. Mean, median, standard deviation, and quartiles still matter. The chart gives those numbers context. A mean can look reasonable even when the data have two clusters, heavy skew, or one bad measurement pulling the scale.
Histogram vs bar chart
Histograms and bar charts both use bars, but the x-axis means different things. A histogram shows ordered numeric ranges such as 0-10, 10-20, and 20-30, and the bars usually touch because the intervals are continuous. A bar chart compares named categories such as products, departments, countries, or survey choices.
Use a histogram when the question is about a distribution: spread, clusters, skew, gaps, or extreme observations. Use a bar chart when the question is about categories: which group is larger, smaller, or changing over time.
A quick check is to ask whether the x-axis order can be rearranged. If changing the order breaks the meaning, you probably have numeric intervals and need a histogram. If the order is mostly a presentation choice, a bar chart is usually the better fit.
How to choose bin size
No bin size works for every histogram. Too few bins can hide structure. Too many bins can make random noise look meaningful. Freedman-Diaconis is a good starting point for skewed data and outliers because it uses the interquartile range. Scott often works better when the data are closer to normal. Sturges is simple and often fine for smaller samples.
For exploration, compare several rules and choose the chart that represents the data honestly. If a class, journal, or team expects a specific method, use that method and name it in the caption. For exact control, set a custom bin count or width.
When two bin settings both look reasonable, prefer the one that is easier to explain. Readers should be able to understand why the intervals were chosen. If several histograms will be compared side by side, use the same bin width and range across all of them.
How to export a histogram for papers and reports
For journals, use SVG or PDF when the submission system accepts vector files. Text, axes, and bars stay sharp when the figure is resized. If the journal requires raster art, export PNG at 300 DPI or higher and choose dimensions that match the target column width. Single-column figures are often around 85 mm wide; double-column figures are often around 180 mm wide.
For slides, 1920 x 1080 px is easy to place in a 16:9 deck. For posters or large reports, use larger pixel dimensions and check the labels at the final display size. You can also export the frequency table as CSV or save the chart settings as JSON.
Before sharing the figure, view it at the size your readers will see. Axis labels, tick labels, legends, and captions should be readable without zooming. If the chart may be printed in grayscale, use a grayscale or high-contrast theme.
Keep the exported table when the exact counts matter. A chart is easier to read, but the table is easier to audit. Saving the JSON settings can also save time later if you need to revise labels, change the DPI, or regenerate the same figure from updated data.