Creating DataFrames

What is a DataFrame?

A pandas DataFrame is a two-dimensional labeled data structure, think of it as a spreadsheet, SQL table, or dictionary of Series objects. Each column is a Series, and all columns share the same index.

import pandas as pd

df = pd.DataFrame({
    "Name": ["Alice", "Bob", "Carol"],
    "Age": [24, 30, 28],
    "Score": [88, 92, 79]
})
print(df)

Output:

    Name  Age  Score
0  Alice   24     88
1    Bob   30     92
2  Carol   28     79

Creating DataFrames from different sources

From a dictionary of lists, each key becomes a column name:

df = pd.DataFrame({
    "City": ["Lagos", "Nairobi", "Cairo"],
    "Population": [15_400_000, 4_400_000, 20_900_000],
    "Country": ["Nigeria", "Kenya", "Egypt"]
})

From a list of dictionaries, each dictionary is a row:

records = [
    {"Name": "Alice", "Score": 88},
    {"Name": "Bob", "Score": 92},
    {"Name": "Carol", "Score": 79},
]
df = pd.DataFrame(records)

From a Series, multiple Series combine into columns:

names = pd.Series(["Alice", "Bob", "Carol"])
scores = pd.Series([88, 92, 79])
df = pd.DataFrame({"Name": names, "Score": scores})

Reading CSV files

The most common way to load real data is from a CSV file:

df = pd.read_csv("titanic.csv")

Useful parameters for read_csv():

df = pd.read_csv(
    "data.csv",
    index_col="id",        # use 'id' column as the index
    usecols=["name", "age", "fare"],  # load only these columns
    na_values=["?", ""],   # treat '?' and empty strings as NaN
    dtype={"age": "float"} # force column type
)

For this course we will use the Titanic dataset, which is available at:

# titanic.csv ships with the course — load it from the browser file system.
df = pd.read_csv("titanic.csv")

Inspecting your data

After loading data, always inspect it first:

df.head()        # first 5 rows
df.tail(3)       # last 3 rows
df.shape          # (rows, columns) — e.g. (891, 12)
df.info()         # column names, non-null counts, dtypes
df.describe()     # statistical summary of numeric columns

The info() method is particularly important, it reveals missing values and data types:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 891 entries, 0 to 890
Data columns (total 12 columns):
 #   Column    Non-Null Count  Dtype  
---  ------    --------------  -----  
 0   PassengerId  891 non-null   int64  
 1   Survived     891 non-null   int64  
 2   Pclass       891 non-null   int64  
 3   Name         891 non-null   object 
 4   Age          714 non-null   float64
 5   SibSp        891 non-null   int64  
 ...

Notice that Age has 714 non-null values out of 891, that means 177 missing values. Cleaning this up is a core skill you will learn later.

Column access

Once you have a DataFrame, you can access columns as Series:

print(df["Age"])       # returns a Series
print(df.Age)          # dot notation also works (if column name has no spaces)

Select multiple columns by passing a list:

df[["Name", "Age"]]

Try It

Create a DataFrame representing three employees with columns for Name, Department, and Salary. Print the DataFrame, then display only the Name and Salary columns.

import pandas as pd

employees = pd.DataFrame({
    "Name": ["Amina", "Kofi", "Zara"],
    "Department": ["Engineering", "Marketing", "Engineering"],
    "Salary": [95000, 72000, 88000]
})
print(employees)
print(employees[["Name", "Salary"]])

Key Takeaways

• A DataFrame is a table with labeled rows (index) and labeled columns
• Dictionaries of lists and lists of dictionaries are the most common construction methods
• pd.read_csv() loads external data, use index_col, usecols, and na_values for control
• Always inspect new data with head(), info(), and describe() before analyzing

Practice Challenge

Load the Titanic dataset from the URL above. How many rows and columns does it have? What are the column names? How many columns have missing values? Use info() to find out.