update to seaborn 0.11.1

README.md

@@ -22,5 +22,24 @@ If I were using Windows, it might look like this:
 Set these aside for now and we'll pick them up in chapter 2.
 
 ## Changelog
+### v0.9.0 (2021-06-16)
+Updated visualization section + associated homework problems to use Seaborn
+0.11.x (September 2020), which added a new `displot` function. This means
+making our distribution plots change from, say:
+
+```python
+g = (sns.FacetGrid(df)
+     .map(sns.kdeplot, 'std', shade=True))
+```
+
+To:
+
+```python
+g = sns.displot(df, x='std', kind='kde', fill=True)
+```
+
+It also opens up some new possibilities (e.g. with plotting empirical CDFs)
+that I might discuss in a future update.
+
 ### v0.8.0
 Add this changelog, bundle files in an github release vs including with SendOwl.

code/06_summary.py

@@ -7,8 +7,7 @@
 # stored
 # on Windows it might be something like 'C:/mydir'
 
-DATA_DIR = '/Users/nathanbraun/fantasymath/fantasybook/data'
-# DATA_DIR = '/Users/nathan/fantasybook/data'
+DATA_DIR = '/Users/nathan/fantasybook/data'
 
 ###############
 # distributions
@@ -37,23 +36,15 @@
 df[['player_name', 'week', 'std', 'ppr', 'half_ppr']].head()
 
 # density plot of standard points
-
-# all on one line
-g = sns.FacetGrid(df).map(sns.kdeplot, 'std', shade=True)
-
-# on seperate lines so it's clearer it's a two step process
-g = (sns.FacetGrid(df)
-     .map(sns.kdeplot, 'std', shade=True))
+g = sns.displot(df, x='std', kind='kde', fill=True)
 
 # density plot of standard points by position
-g = (sns.FacetGrid(df, hue='pos')
-     .map(sns.kdeplot, 'std', shade=True)
-     .add_legend())
+g = sns.displot(df, x='std', kind='kde', fill=True, hue='pos')
+
 
 # density plot of standard points by position and week
-g = (sns.FacetGrid(df, hue='pos', col='week', col_wrap=4, height=2)
-     .map(sns.kdeplot, 'std', shade=True)
-     .add_legend())
+g = sns.displot(df, x='std', kind='kde', fill=True, hue='pos', col='week',
+                col_wrap=4, height=2)
 
 # now want it by scoring system
 
@@ -73,15 +64,28 @@ def score_type_df(_df, scoring):
     [score_type_df(df, scoring) for scoring in ['std', 'ppr', 'half_ppr']])
 
 # now can plot points by scoring system and position
-g = (sns.FacetGrid(df_pts_long, col='pos', hue='scoring', col_wrap=2,
-                   aspect=1.3)
-     .map(sns.kdeplot, 'pts', shade=True)
+g = sns.displot(df_pts_long,  x='pts', col='pos', hue='scoring', kind='kde',
+                fill=True, col_wrap=2, aspect=1.3)
+
+# ecdf
+g = (sns.displot(df_pts_long,  x='pts', col='pos', hue='scoring', kind='ecdf',
+                 col_wrap=2, aspect=1.3)
      .add_legend())
+plt.show()
+
 
 #################################
 # relationships between variables
 #################################
 
+
+# airyards vs yac
+g = sns.displot(x='caught_airyards', y='raw_yac', data=df, kind='kde')
+plt.show()
+
+g.fig.subplots_adjust(top=0.9)
+g.fig.suptitle('Airyards vs YAC')
+
 # airyards vs yac
 g = sns.relplot(x='caught_airyards', y='raw_yac', data=df)
 g.fig.subplots_adjust(top=0.9)
@@ -133,15 +137,22 @@ def score_type_df(_df, scoring):
 ##############
 
 # basic plot
-g = (sns.FacetGrid(df_pts_long, col='pos', hue='scoring')
-     .map(sns.kdeplot, 'pts', shade=True))
+g = sns.displot(df_pts_long,  x='pts', col='pos', hue='scoring', kind='kde',
+                fill=True)
 
 # wrapping columns
-g = (sns.FacetGrid(df_pts_long, col='pos', hue='scoring', col_wrap=2)
-     .map(sns.kdeplot, 'pts', shade=True)
-     .fig.subplots_adjust(top=0.9) # adding a title
-     .fig.suptitle('Fantasy Points by Position, Scoring System')
-     .set_xlabels('Points') # modifying axis
-     .set_ylabels('Density')
-     .add_legend()  # adding legend
-     .savefig('scoring_by_pos_type.png'))  # saving
+g = sns.displot(df_pts_long,  x='pts', col='pos', hue='scoring', kind='kde',
+                fill=True, col_wrap=2)
+
+# adding a title
+g.fig.subplots_adjust(top=0.9) # adding a title
+g.fig.suptitle('Fantasy Points by Position, Scoring System')
+
+# modifying axis
+g.set(xlim=(-10, 40), ylim=(0, 0.014))
+
+g.set_xlabels('Points') # modifying axis
+g.set_ylabels('Density')
+
+# saving
+g.savefig('scoring_by_pos_type.png')  # saving

code/07_02_coinflip.py

@@ -2,7 +2,6 @@
 from pandas import DataFrame
 import statsmodels.formula.api as smf
 
-
 coin = ['H', 'T']
 
 # make empty DataFrame

solutions-to-exercises/06_plotting_answers.py

@@ -11,57 +11,54 @@
 # stored
 # on Windows it might be something like 'C:/mydir'
 
-DATA_DIR = '/Users/nathan/fantasybook/data'
+# DATA_DIR = '/Users/nathan/fantasybook/data'
+DATA_DIR = '/Users/nathanbraun/fantasymath/ltcwff-files/data'
 
 pbp = pd.read_csv(path.join(DATA_DIR, 'play_data_sample.csv'))  # play by play data
 
 ###############################################################################
 # 6.1a
 ###############################################################################
-g = (sns.FacetGrid(pbp)
-     .map(sns.kdeplot, 'yards_gained', shade=True))
+g = sns.displot(pbp, x='yards_gained', kind='kde', fill=True)
 g.fig.subplots_adjust(top=0.9)
 g.fig.suptitle('Distribution of Yards Gained Per Play, LTCWFF Sample')
-g.savefig('./answers-to-exercises/6-1a.png')
+g.savefig('./solutions-to-exercises/6-1a.png')
 
 ###############################################################################
 # 6.1b
 ###############################################################################
-g = (sns.FacetGrid(pbp.query("down <= 3"), hue='down')
-     .map(sns.kdeplot, 'yards_gained', shade=True))
-g.add_legend()
+g = sns.displot(pbp.query("down <= 3"), x='yards_gained', kind='kde',
+                hue='down', fill=True)
 g.fig.subplots_adjust(top=0.9)
 g.fig.suptitle('Distribution of Yards Gained Per Play by Down, LTCWFF Sample')
-g.savefig('./answers-to-exercises/6-1b.png')
+g.savefig('./solutions-to-exercises/6-1b.png')
 
 ###############################################################################
 # 6.1c
 ###############################################################################
-g = (sns.FacetGrid(pbp.query("down <= 3"), col='down')
-     .map(sns.kdeplot, 'yards_gained', shade=True))
+g = sns.displot(pbp.query("down <= 3"), x='yards_gained', kind='kde',
+                col='down', fill=True)
 g.fig.subplots_adjust(top=0.8)
 g.fig.suptitle('Distribution of Yards Gained Per Play by Down, LTCWFF Sample')
-g.savefig('./answers-to-exercises/6-1c.png')
+g.savefig('./solutions-to-exercises/6-1c.png')
 
 ###############################################################################
 # 6.1d
 ###############################################################################
-g = (sns.FacetGrid(pbp.query("down <= 3"), col='down', row='posteam')
-     .map(sns.kdeplot, 'yards_gained', shade=True))
+g = sns.displot(pbp.query("down <= 3"), x='yards_gained', kind='kde',
+                row='posteam', col='down', fill=True)
 g.fig.subplots_adjust(top=0.9)
 g.fig.suptitle('Distribution of Yards Gained Per Play by Down, Team, LTCWFF Sample')
-g.savefig('./answers-to-exercises/6-1d.png')
+g.savefig('./solutions-to-exercises/6-1d.png')
 
 ###############################################################################
 # 6.1e
 ###############################################################################
-g = (sns.FacetGrid(pbp.query("down <= 3"), col='down', row='posteam',
-                   hue='posteam')
-     .map(sns.kdeplot, 'yards_gained', shade=True))
+g = sns.displot(pbp.query("down <= 3"), x='yards_gained', col='down',
+                row='posteam', hue='posteam', fill=True)
 g.fig.subplots_adjust(top=0.9)
 g.fig.suptitle('Distribution of Yards Gained Per Play by Down, Team, LTCWFF Sample')
-g.savefig('./answers-to-exercises/6-1e.png')
-
+g.savefig('./solutions-to-exercises/6-1e.png')
 
 ###############################################################################
 # 6.2