Question $1\mathrm{.}2.$ Now it's time to count the number of episodes each director directed. Make a DataFrame called directors, indexed by 'name', with one column called 'num$\_$episodes'. There should be one row for each unique director, and 'num$\_$episodes' should contain the number of episodes they directed or co$-$directed. Sort this DataFrame so that the most experienced directors appear at the top.

Hints:

Our solution involved creating a new, empty DataFrame with bpd$.$DataFrame$(),$ adding columns, and using groupby.

For groupby to give meaningful results, you must use it on a DataFrame with at least two columns.

directors$\_$by$\_$episode $=$ episodes$[\text{'}$directed$\_$by$\text{'}].$apply$($parse$\_$names$)$

all$\_$directors $=[$director for directors in directors$\_$by$\_$episode for director in directors$]$

directors$\_$df $=$ pd$.$DataFrame$($all$\_$directors, columns$=[\text{'}$name$\text{'}])$

director$\_$counts $=$ directors$\_$df$[\text{'}$name$\text{'}].$value$\_$counts$().$reset$\_$index$()$

director$\_$counts.columns $=[\text{'}$name$\text{'},$ 'num$\_$episodes'$]$

directors $=$ director$\_$counts.sort$\_$values$($by$=$'num$\_$episodes', ascending$=$False$).$reset$\_$index$($drop$=$True$)$

directors

directors$\_$by$\_$episode $=$ episodes$[\text{'}$directed$\_$by$\text{'}].$apply$($parse$\_$names$)$

$$

all$\_$directors $=[$director for directors in directors$\_$by$\_$episode for director in directors$]$

$$

directors$\_$df $=$ pd$.$DataFrame$($all$\_$directors, columns$=[\text{'}$name$\text{'}])$

$$

director$\_$counts $=$ directors$\_$df$[\text{'}$name$\text{'}].$value$\_$counts$().$reset$\_$index$()$

director$\_$counts.columns $=[\text{'}$name$\text{'},$ 'num$\_$episodes'$]$

$$

directors $=$ director$\_$counts.sort$\_$values$($by$=$'num$\_$episodes', ascending$=$False$).$reset$\_$index$($drop$=$True$)$

directors

name num$\_$episodes

$0$ James Burrows $3$

$1$ Kevin S$.$ Bright $3$

$2$ Gary Halvorson $2$

grader.check$("$q$1\_2")$

q$1\_2$ results:

q$1\_2-1$ result:

Trying:

isinstance$($directors$,$ bpd$.$DataFrame$)$ and directors.shape $==(29,1)$ # Check number of rows and columns

Expecting:

True

$**********************************************************************$

Line $1,$ in q$1\_20$

Failed example:

isinstance$($directors$,$ bpd$.$DataFrame$)$ and directors.shape $==(29,1)$ # Check number of rows and columns

Expected:

True

Got:

False

Question $1\mathrm{.}1.$ We want to first handle any episodes with two directors so that later we can count the number of episodes each director contributed to$.$ Create a function called parse$\_$names that turns a string in the 'directed$\_$by$\text{'}$ column to a list of one or two names. Values in the 'directed$\_$by$\text{'}$ column are either single names, or two names separated by an ampersand. For example,

parse$\_$names$(\text{'}$James Burrows'$)$ should return $[\text{'}$James Burrows'$],$ and

parse$\_$names$(\text{'}$Kevin S$.$ Bright & Gary Halvorson'$)$ should return $[\text{'}$Kevin S$.$ Bright', 'Gary Halvorson'$].$

Then, apply your function to the 'directed$\_$by$\text{'}$ column from the episodes DataFrame, and store the resulting Series in the variable directors$\_$by$\_$episode. Do not modify the episodes DataFrame.For the next question, you'll need to know something interesting about how lists work in Python: when you sum two lists together, the output is one giant list that contains all the elements in both lists combined. An example is shown below.

In $[10]$: $[\text{'}$List$\text{'},$ 'combining'$]+[\text{'}$is$\text{'},\text{'}$my$\text{'},$ 'passion'$]$

Logout

Control $P_{c^{}}$

Not Trusted

Python $3($ipykern

Run

Markdown

Validate

LHI

$-$

git nbdiff

Out$[10]$: $[\text{'}$List$\text{'},$ 'combining', $\text{'}$is$\text{'},\text{'}$my$\text{'},$ 'passion'$]$

Question $1\mathrm{.}2.$ ~~ Now it's time to count the number of episodes each director directed. Make a DataFrame called directors $,$ indexed by $\text{'}$ name' with one column called 'num$\_$episodes'. There should be one row for each unique director, and 'num$\_$episodes' should contain the number of episodes they directed or co$-$directed. Sort this DataFrame so that the most experienced directors appear at the top.

Hints:

Our solution involved creating a new, empty DataFrame with bpd$.$ DataFrame$(),$ adding columns, and using groupby $.$

For groupby to give meaningful results, you must use it on a DataFrame with at least two columns.