Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsTo create a useful chart in Python, start with the data question, identify the variables and their order, then choose a plot that matches the comparison you need. pandas is the quickest route from a Series or DataFrame to a familiar chart; seaborn adds statistical groupings and facets; and Matplotlib provides direct control over figures and axes. These tools work together rather than being mutually exclusive.
Start with the question, not the library
A chart is a visual answer to a data question. Before writing plotting code, decide whether you need to show change, a relationship, category differences, a distribution, or group spread. Record the units, category names, time range, sample size, and whether each value is a raw observation or an aggregation.
Choose a chart by data structure
| Question | Good starting chart | What to check |
|---|---|---|
| How does a value change along an ordered axis or through time? | Line plot | The x-values have meaningful order; lines imply continuity. |
| How are two numeric variables related? | Scatter plot | Look for overlap, clusters, nonlinearity, and possible outliers. |
| How do named categories compare? | Bar plot | Use a common baseline and label the measured unit. |
| How are values distributed? | Histogram | Bin width can change the apparent shape; consider an ECDF or KDE when appropriate. |
| How do groups differ in spread and possible outliers? | Box plot | Show raw points as well when sample size or multimodality matters. |
| Do several groups or variables need separate views? | Facets or small multiples | Keep scales and encodings comparable across panels. |
These are introductory defaults, not rules. Measurement scale, overlap, aggregation, uncertainty, and sample size can justify a different design. OpenStax’s data-visualization chapter distinguishes histograms for continuous-variable distributions, box plots for quartiles and possible outliers, and line plots for trends over time.
A practical table-to-chart workflow
- Prepare a small, understandable table. Each column should have a clear meaning, consistent units, and suitable missing-value handling.
- State the question and variable roles. Identify which field is the ordered or categorical x variable, which is numeric, and which fields define groups.
- Choose a plot family. Use the chart table above as a starting point.
- Map variables explicitly. In a plotting call, name roles such as
x,y,hue, and facet columns rather than relying on implicit column order. - Label the result. Include a descriptive title when useful, axis labels with units, readable category names, and a stated time range.
- Check what the marks represent. A mean, median, confidence interval, fitted line, or smoothed density is not the same as an individual observation.
- Customize for the audience. Adjust scales, tick labels, colors, annotations, and layout only when they improve interpretation.
- Save or share the figure. Export the finished figure at a size and format appropriate for its destination.
Plot a pandas DataFrame quickly
Series.plot and DataFrame.plot are low-friction entry points for common charts. Pandas documents line, area, bar, horizontal bar, box, density, hexbin, histogram, KDE, pie, and scatter methods in its chart-visualization guide and beginner plotting tutorial. A column is normally drawn as a separate visual element; subplots=True can place columns in separate panels.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →import pandas as pd
# date and value are columns in the table
df = pd.read_csv("measurements.csv", parse_dates=["date"])
df.plot(x="date", y="value")This is convenient for exploration, but it does not decide whether the chart answers your question well. Sort an ordered x column when necessary, verify units, and inspect missing or duplicated observations before interpreting the line.
Keep pandas convenience and Matplotlib control
Pandas plotting methods return Matplotlib objects. You can prepare an Axes, let pandas draw into it, then use Matplotlib for labels, layout, annotations, and output.
Rank #2
import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(8, 4))
df.plot(x="date", y="value", ax=ax, legend=False)
ax.set_xlabel("Date")
ax.set_ylabel("Value (units)")
ax.set_title("Value over time")
fig.tight_layout()
fig.savefig("chart.png", dpi=160)Use Matplotlib directly when pandas does not expose the plot type or customization you need. Its official plot-type guide covers pairwise, distribution, gridded, irregular-grid, and 3D or volumetric functions. For beginners, line, scatter, bar, histogram, and box plots cover most first questions.
Use seaborn for statistical and grouped views
Seaborn is a higher-level interface for relational, distributional, categorical, estimation, regression, and multi-view graphics. Its functions accept pandas or NumPy objects and, where supported, Python lists and dictionaries. It is especially useful when color, style, size, rows, or columns encode groups.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Rank #3
For long-form data, put one variable in each column and one observation in each row. Then map columns to visual roles explicitly:
import seaborn as sns
sns.relplot(
data=df,
x="year",
y="passengers",
hue="month",
kind="line"
)The seaborn data-structure guide uses this kind of mapping with its flights example. Long-form organization makes grouping and faceting explicit; support for particular input forms can differ between seaborn functions, so check the function’s documentation when passing non-tabular data.
Understand estimates, smoothing, and uncertainty
A chart can display raw observations or summarize them. A bar may show a total or an average; a line may connect observations or show an estimate; a regression line and a KDE are fitted or smoothed views. State the aggregation or statistical operation in the label or accompanying text, and show uncertainty when it matters.
Seaborn separates statistical estimation, error bars, regression fits, and distribution visualization in its guide. Do not let a confidence interval, error bar, or smooth curve look like additional measured data. For distributions, explain important bin or smoothing choices, particularly when different settings could lead readers to different conclusions.
Best Value
Which Python library should you use?
| Library | Best starting point | Typical reason to move beyond it |
|---|---|---|
| pandas | Fast exploratory charts directly from a Series or DataFrame. |
You need statistical semantics, richer grouping, or a plot method pandas does not provide. |
| seaborn | Grouped statistical graphics, relational and categorical views, and facets. | You need lower-level construction or specialized Matplotlib control. |
| Matplotlib | Direct construction and detailed control of figures, axes, scales, labels, and output. | You want a higher-level interface for routine table mappings. |
These interfaces overlap. A common pattern is to prepare data with pandas, draw a grouped view with seaborn, and use Matplotlib objects for final formatting and saving.
Checks before you publish or present a chart
- Does the chosen mark match the variable types and the question?
- Are ordered values actually ordered, and are categories clearly named?
- Are axes, units, zero baselines, and time ranges understandable?
- Could overplotting, too many categories, or a misleading scale hide the pattern?
- Does the figure distinguish observations from averages, fitted values, and uncertainty?
- Would a raw-point overlay, facet, annotation, or table make an important difference easier to inspect?
- Can the intended audience read the colors, labels, and text at the final display size?
Documentation and version awareness
Library APIs change. The documentation consulted for this introduction identifies pandas 3.0.6, seaborn 0.13.2, and Matplotlib 3.11.0; verify the live official documentation before relying on version-specific behavior or code. For broader Python study, Introduction to Python Programming by Udayan Das, Aubrey Lawson, Chris Mayfield, and Narges Norouzi (OpenStax, published March 13, 2024) includes a data-visualization chapter, but it is a general introductory textbook rather than a dedicated visualization reference.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
