Share your ASGN 3 figures from Q10 here!

[donbowen]

@LeDataSciFi/classmates-2024

[elvinlee12]

[justinreed23]

Here's my plot for Q10!

[JinyiXu]

[Keith-Cheung]

[hanadinh]

[dannyhubert9]

[NickJS1231]

[adrianmross]

[donbowen]

Q for next class @adrianmross - why so many 0s?

Q from @dannyhubert9

[bmnguyen6403]

[benciancio]

[kevinchen024]

[evandeblase]

[DannyPress]

[nol223]

[donbowen]

I'm guessing you did some kind of reshape of the data before this?

[MariaMaragkelli]

[nicoschuster01]

[donbowen]

The diagonals will always be 1, and thus "overwhelm" the color scale - so it's nice to drop them. And the upper part of the triangle is redundant. So on https://ledatascifi.github.io/ledatascifi-2024/content/03/04e-visualEDA.html?highlight=heatmap, I add thsi "mask"" argument, which does that:

# dont plot identifying type info or categorical vars
corr = ccm.drop(columns=['gvkey','lpermno','sic3','fyear','sic']).corr()

fig, ax = plt.subplots(figsize=(9,9)) # make a big space for the figure
ax = sns.heatmap(corr,
                 # cmap for the colors, 
                 center=0,square=True,
                 cmap=sns.diverging_palette(230, 20, as_cmap=True),
                 # mask to hide the upper diag (redundant)
                 mask=np.triu(np.ones_like(corr, dtype=bool)),
                 # shrink the heat legend
                 cbar_kws={"shrink": .5},
                 #optional: vmax and vmin will "cap" the color range
                )

[alz425]

[dhrunal13]

[acg425]

[Danielshin2002]

not 100% sure why the other gsectors have so many null values, was planning to draw the efficient frontier..

[michaelparker7]

[chrisztoh]

[samfleet]

[keishamodi]

[donbowen]

kind = hex, right?

[leosc326]

[masonschick]

[Saintwy6]

[donbowen]

NOTE: This code was added by the professor in advance of Assignment 4. Assignment code should not be posted by students like this.

# Q10 - Cumulative returns by industry and leverage

g = sns.relplot(
    data=df_new,
    x='week', y='cum_ret$', col="portfolio", hue="portfolio",
    kind="line", palette="crest", linewidth=4, zorder=5,
    col_wrap=2, height=2, aspect=2, legend=False,
)

# Iterate over each subplot to customize further
for year, ax in g.axes_dict.items():

    # Add the title as an annotation within the plot
    ax.text(.45, .85, year, transform=ax.transAxes, fontweight="bold")

    # Plot every year's time series in the background
    sns.lineplot(
        data=df_new, x='week', y='cum_ret$', units="portfolio",
        estimator=None, color=".7", linewidth=1, ax=ax,
    )

# Reduce the frequency of the x axis ticks
ax.set_xticks(ax.get_xticks()[::2])

# Tweak the supporting aspects of the plot
g.set_titles("")
g.set_axis_labels("Weeks", "Returns")
g.fig.suptitle('Portfolio Returns on $100 (Feb, Mar, Apr 2020)')
g.fig.subplots_adjust(top = 0.85)

[cillianfisher]

[acw224-15]