Python — Pandas

Install & the two core types

pip install pandas

import pandas as pd

# Series — one labeled column
s = pd.Series([90, 85, 77], name="score")

# DataFrame — a full table (the star of the show)
df = pd.DataFrame({
    "name":  ["Ada", "Bo", "Cleo"],
    "age":   [36, 22, 45],
    "score": [91, 78, 88],
})

   name  age  score
0   Ada   36     91      ← the numbers on the left are the index
1    Bo   22     78
2  Cleo   45     88

Loading real data

df = pd.read_csv("users.csv")
df = pd.read_excel("sales.xlsx")       # needs openpyxl installed
df = pd.read_json("data.json")

df.head()        # first 5 rows — your first command every time
df.shape         # (rows, columns)
df.info()        # dtypes + missing-value overview
df.describe()    # instant statistics summary

Selecting columns & rows

df["name"]                  # one column → Series
df[["name", "score"]]       # multiple → DataFrame

df.loc[2]                   # row by LABEL
df.iloc[0]                  # row by POSITION

df.at[1, "age"]             # single cell: label row, column name

Filtering — boolean masks

df[df["score"] > 80]                    # passing students
df[(df["age"] < 30) & (df["score"] > 80)]   # AND — parentheses required!
df[(df["role"] == "admin") | df["active"]]

Adding & changing columns

df["passed"] = df["score"] >= 80            # computed column
df["score_adj"] = df["score"] + 5           # vectorized like NumPy
df = df.rename(columns={"name": "full_name"})
df = df.drop("temp", axis=1)                # remove column

Grouping & aggregation

df.groupby("department")["salary"].mean()
df.groupby("dept").agg(
    avg=("salary", "mean"),
    count=("salary", "count"),
)
df.sort_values("score", ascending=False)
df["score"].value_counts()

Handling missing data

df.isna().sum()                     # how many NaNs per column
df.dropna()                         # drop incomplete rows
df.fillna(0)                        # fill with value
df["age"].fillna(df["age"].median())    # smarter fill

The typical analysis in 6 lines

import pandas as pd

df = pd.read_csv("sales.csv")
monthly = df.groupby("month")["revenue"].sum()
best = monthly.idxmax()

print(f"Best month: {best} (${monthly.max():,.0f})")

Gotchas: chained indexing (df[a][b] = x) silently fails to write — use .loc; &/| not and/or inside filters; SettingWithCopyWarning means operate on .copy().

Mini Practice

1. Build a 5-row DataFrame by hand; run head/describe.
2. Filter rows by two conditions with & .
3. Add a derived column from arithmetic on others.
4. groupby category → mean of values; sort result.
5. Load any CSV; report shape, missing counts, top row.

Next: matplotlib →