Pandas Data Manipulation: Sorting, Selecting, Grouping, and Reshaping Techniques

Pandas is a widely used tool for data mining, and mastering its functions is essential; this article visually demonstrates a variety of Pandas operations.

sort_values : Filter rows where the size column equals "medium", sort by type , then group by type and compute the median. (dogs[dogs['size'] == 'medium'] .sort_values('type') .groupby('type').median())

Selecting a column : Access a single column directly, e.g., dogs['longevity'] .

groupby + mean : Group data by size and calculate the mean of each group. dogs.groupby('size').mean()

Grouping multiple columns : Perform a groupby on both type and size . dogs.groupby(['type', 'size'])

groupby + multi aggregation : Sort by size , group by size , then aggregate the height column with sum, mean, and standard deviation. (dogs.sort_values('size') .groupby('size')['height'].agg(['sum', 'mean', 'std']))

Filtering for columns : Use df.loc with a condition to select specific columns, e.g., df.loc[:, df.loc['two'] <= 20] .

Filtering for rows : Combine multiple conditions to filter rows and select a column, e.g., dogs.loc[(dogs['size'] == 'medium') & (dogs['longevity'] > 12), 'breed'] .

Dropping columns : Remove unwanted columns with dogs.drop(columns=['type']) .

Joining : Join two DataFrames using ppl.join(dogs) .

Merging : Merge DataFrames on specific keys with a left join. ppl.merge(dogs, left_on='likes', right_on='breed', how='left')

Pivot table : Create a pivot table to summarize data. dogs.pivot_table(index='size', columns='kids', values='price')

Melting : Transform wide format to long format using dogs.melt() .

Pivoting : Pivot the DataFrame with specified index and columns. dogs.pivot(index='size', columns='kids')

Stacking column index : Stack the column index into a Series with dogs.stack() .

Unstacking row index : Unstack the row index back to a DataFrame using dogs.unstack() .

Resetting index : Reset the index to default integer indexing via dogs.reset_index() .

Setting index : Set a specific column as the index, e.g., dogs.set_index('breed') .