3.3 Built-In Pandas Functions that work on Dataframes

Question 1

What built-in pandas function lets you take the mean of numeric columns in a dataframe?

Answer 1

.mean()

pg[[‘shot’, ‘goal’, ‘assist’]].mean()
–> gives the mean of each column

Question 2

What built-in pandas function lets you take the max of numeric columns in a dataframe?

Answer 2

.max()

pg[[‘shot’, ‘goal’, ‘assist’]].max()
–> gives the max of each column

Question 3

What does max() do that mean() doesn’t?

Answer 3

Max() can also operate on string columns as it treats the “max” as the latest in the alphabet

Question 4

What is important to remember about axis when using pandas built-in functions?

Answer 4

Whether you want to calculate the statistic on the columns or the rows:

columns (the default, axis=0),
rows (axis=1)

Question 5

Say you want to calculate the mean number of goals by a player, and the player data is spread out per match (7 columns for goals, 1 column representative per match), what should the axis argument equal?

Answer 5

axis=1, as the mean should be calculated across the rows, and NOT columns!

Using axis=0 would give the mean goals scored per game by ALL players

Question 6

What do 0 and 1 mean when using built-in summary stats on boolean columns?

Answer 6

0 = False
1 = True

Question 7

What are the two boolean specific summary functions (Pandas)?

Answer 7

.any()

.all()

Question 8

What does the .any() boolean specific function do?

Answer 8

eg. Can be used to check is any player scored above 100 passes

(pg[‘pass’] > 100).any()

Question 9

What does the .all() boolean specific function do?

Answer 9

eg. Can be used to check is ALL players made at least one pass

(pg[‘pass’] > 0).all()

Question 10

Do the boolean specific functions “any()” and “all()” take an argument?

Answer 10

Yes, they take axis arguments.

Could set the axis to be 1, this will go through each row (player) and check if they won more than 5 air duels

Question 11

Which function returns a summary of the frequency of individual values?

Answer 11

.value_counts()

Question 12

What argument can be used in addition to the “.value_counts()” function which ensures the values will add up to 1 and represent propotions?

Answer 12

(normalize= True)

pg[‘team’].value_counts(normalize=True)

–> this grabs the column of teams and outputs a % for the frequency of a team name appearing in the datalist.

Question 13

What does the “crosstab” function do?

Answer 13

Shows the frequencies for ALL the combinations of the two columns

pd.crosstab(pg[‘team’], pg[‘pos’])

–> this outputs a frame showing how often a country (team) is present in the datafile in combination with positions

(eg. ‘romelu lukaku’, ‘belgium’, ‘FWD’)
–> This is 1 instance of the position ‘FWD’ and ‘Belgium’ being combined

Question 14

Which pandas function is used to change the dataframe format from wide to long?

Answer 14

melt() and unmelt()

“Unpivot a DataFrame from wide to long format, optionally leaving identifiers set.”

Question 15

Which pandas function allows you to restructure a dataframe by turning rows into columns?

Answer 15

pivot()

“Return reshaped DataFrame organized by given index / column values.”

Question 16

Which pandas function allows you to combine DataFrames or named Series into a singular DataFrame?

Answer 16

.merge()

“Merge DataFrame or named Series objects with a database-style join.”

Question 17

Which pandas function attempts to infer the datetime format?

Answer 17

.to_datetime()

“Pandas to_datetime will attempt to infer the datetime format based on the input, which can make parsing faster.”