Altair Basics & Chart Object

📦 Install

pip install altair vega_datasets

import altair as alt
import pandas as pd
from vega_datasets import data      # built-in sample datasets

🧱 The core pattern

alt.Chart(df).mark_point().encode(
    x="col1",
    y="col2"
)

Piece	Role
Chart(df)	wraps a pandas DataFrame
.mark_X()	how to draw each row — point, bar, line, etc.
.encode(...)	which columns map to which visual channels (x, y, color…)

Nothing renders until .encode() is called — mark_point() alone just sets the shape, encode() is what actually connects data to the chart.

📊 Sample datasets

df = data.cars()
df = data.iris()
df = data.stocks()

👁️ Displaying a chart

chart = alt.Chart(df).mark_bar().encode(x="a", y="b")
chart                    # Jupyter: auto-displays last expression in a cell
chart.show()                 # opens in browser, works outside Jupyter too

🔤 Shorthand vs explicit encoding

# shorthand — column name as a plain string, Altair infers the type
alt.Chart(df).mark_bar().encode(x="category", y="value")
 
# explicit — wrap in alt.X()/alt.Y() for full control
alt.Chart(df).mark_bar().encode(
    x=alt.X("category", type="nominal", title="Category"),
    y=alt.Y("value", type="quantitative", title="Value")
)

Start with shorthand strings for speed. Switch to alt.X(...)/alt.Y(...) the moment you need to set a title, sort order, scale, or explicit type. See Altair Data Types & Transformations for the type shorthand codes (:Q, :N, etc).

🔗 Type shorthand suffix

alt.Chart(df).mark_bar().encode(
    x="category:N",     # nominal
    y="value:Q"             # quantitative
)

Covered fully in Altair Data Types & Transformations.

📐 Chart size

alt.Chart(df).mark_point().encode(x="a", y="b").properties(width=400, height=300)

🏷️ Title

alt.Chart(df).mark_bar().encode(x="a", y="b").properties(title="My Chart")
alt.Chart(df).mark_bar().encode(x="a", y="b").properties(
    title=alt.TitleParams("My Chart", subtitle="A subtitle here")
)

🔗 Next

Altair Marks & Encodings · Altair Data Types & Transformations