CDIPpy Example - Boxplot of Monthly-Averaged Significant Wave Height¶

##- Get Station Dataset object
stn_data = StnData(stn)

##- Get metadata (i.e. information about individual deployments)
meta = stn_data.get_stn_meta()
stn_name = meta['metaStationName']
print(stn_name)

##- Use CDIPPY to convert input start/end date strings to datetime objects
import cdippy.utils.utils as utils
start = utils.cdip_datetime(sdate)
end = utils.cdip_datetime(edate)

##- Grab data using 'get_series' function
data = stn_data.get_series(start, end, params)
data.keys()

## Create pandas dataframe from data
df = pd.DataFrame({'Time': wT, 'Hs': data['waveHs'].data})
df.set_index('Time', inplace=True)
# Add a "Month" column for grouping
df['Month'] = df.index.month
df['Hs']

# Create figure and specify figure size
fig = plt.figure(figsize=(15,12))

# This pushes the axes down so the buoy title can be prominent
fig.subplots_adjust(top=0.87)

# Use grid spec to define a 30 column grid. Total boxplot uses 4 columns
ax = plt.subplot2grid((1, 30), (0, 0), colspan=25)
ax2 = plt.subplot2grid((1, 30), (0, 26), colspan=4)

# Create another axes for Hs in feet
ax_feet = ax.twinx()
ax_feet.set_ylabel("Hs (ft)", fontsize=label_font_size + 1)

# Get ylim top and set axes ylims
ylim_top = max(default_ylim_top, max(data["waveHs"]))  # meters
ylim = (0, int(ylim_top))
ax.set_ylim(ylim)
ax_feet.set_ylim(ft_to_m(ylim[0]), ft_to_m(ylim[1]))
ax2.set_ylim(ylim)

# Create months boxplot using axes ax. Subplots=False plots on a single axes on the current figure
(ax, bp_dict) = grp.boxplot(
    return_type="both",
    subplots=False,
    showmeans=True,
    meanline=True,
    whiskerprops={"linestyle": "solid"},
    ax=ax,
)

# Set colors
# Could use plt.setp(bp_dict['caps'], color=box_color)
[
    x.set(markeredgecolor=flier_color, markerfacecolor=flier_color)
    for x in bp_dict["fliers"]
]
[x.set(color=box_color) for x in bp_dict["caps"]]
[x.set(color=box_color) for x in bp_dict["whiskers"]]
[x.set(color=box_color) for x in bp_dict["boxes"]]

# Turn off vertical grid
ax.xaxis.grid(False)

# Turn on minor ticks for both y axes. Feels kludgey using plt.sca (set current axes)
plt.sca(ax)
plt.minorticks_on()
plt.sca(ax_feet)
plt.minorticks_on()

# Turn off unwanted ticks. Could also set labelsize here
ax.tick_params(axis="x", which="minor", bottom="off", top="off")
ax.tick_params(axis="x", which="major", top="off")
ax_feet.tick_params(axis="x", which="minor", bottom="off", top="off")

# Loop through months, set colors, avg, x labels and counts
counts = grp.count()
labels = []
total_records = 0
for m in range(1, 13):
    if m in months_with_data:
        mn = bp_dict["means"][m - 1]
        mn.set(color=mean_color)
        x = mn.get_xdata()
        y = mn.get_ydata()
        xy = ((x[0] + x[1]) / 2, y[0] * (1 + 0.1 / ylim[1]))
        # Annotate means
        ax.annotate(
            str(round(y[0], 2)),
            xy=xy,
            color=mean_color,
            ha="center",
            fontsize=annotate_font_size,
        )
        # Create labels w/ record counts
        labels.append(" " + calendar.month_abbr[m] + "\n" + str(counts[0][m]))
        total_records += counts[0][m]
    else:
        labels.append(" " + calendar.month_abbr[m] + "\n" + "0")
ax.set_xticklabels(labels, fontsize=label_font_size)

# Adjust yticklabel font size
plt.setp(ax.get_yticklabels(), fontsize=label_font_size)
plt.setp(ax_feet.get_yticklabels(), fontsize=label_font_size)

# Create subplot of totals for the year
(ax2, bp_dict) = df_totals.boxplot(
    return_type="both",
    showmeans=True,
    meanline=True,
    whiskerprops={"linestyle": "solid"},
    ax=ax2,
)

# Set colors
bp_dict["means"][-1].set(color=mean_color)
[
    x.set(markeredgecolor=flier_color, markerfacecolor=flier_color)
    for x in bp_dict["fliers"]
]
[x.set(color=box_color) for x in bp_dict["caps"]]
[x.set(color=box_color) for x in bp_dict["whiskers"]]
[x.set(color=box_color) for x in bp_dict["boxes"]]

# Turn off the axis
plt.sca(ax2)
plt.axis("off")


# Get x,y coordinates for totals annotations
x = bp_dict["caps"][0].get_xdata()[1]
cap_lower_y = bp_dict["caps"][0].get_ydata()[0]
cap_upper_y = bp_dict["caps"][1].get_ydata()[0]
whisker_lower_y = bp_dict["whiskers"][0].get_ydata()[0]
whisker_upper_y = bp_dict["whiskers"][1].get_ydata()[0]
median_y = bp_dict["medians"][0].get_ydata()[0]
mean_y = bp_dict["means"][0].get_ydata()[0]
whisker_avg = (whisker_upper_y + whisker_lower_y) / 2.0

# Make the annotations
ax2.annotate(
    "-1.5*L",
    xy=(x, cap_lower_y),
    color="m",
    fontsize=annotate_font_size,
     va="center",
)
ax2.annotate("+1.5*L", xy=(x, cap_upper_y), color="m", fontsize=annotate_font_size)
ax2.annotate(
    "25%ile",
    xy=(x, whisker_lower_y),
    color="blue",
    fontsize=annotate_font_size,
    va="top",
)
ax2.annotate(
    "75%ile", xy=(x, whisker_upper_y), color="blue", fontsize=annotate_font_size
)
ax2.annotate(
    str(round(median_y, 1)) + " (median)",
    xy=(x + 0.2, median_y),
    color=median_color,
    fontsize=annotate_font_size,
)
ax2.annotate(
    str(round(mean_y, 1)) + " (mean)",
    xy=(x + 0.2, mean_y),
    color=mean_color,
    fontsize=annotate_font_size,
)
ax2.annotate(
    "L",
    xy=(x - 0.3, whisker_avg),
    color="m",
    fontsize=subtitle_font_size - 1,
    ha="right",
    va="center",
)
ax2.annotate(
    "Total" + "\n" + str(total_records),
    xy=(x, 0),
    xytext=(x, -0.3 / ylim[1]),
    color="k",
    fontsize=subtitle_font_size - 1,
    ha="center",
    va="top",
)

# Add the title and suptitle, adjust a little to the left to center over months boxplot axes
plt.suptitle(stn_name, fontsize=title_font_size, transform=ax.transAxes, x=0.45)
ax.set_title(
    "Significant Wave Height by month for " + str(start.year), fontsize=subtitle_font_size
)

# Adjust ylabel font size
ax.set_ylabel("Significant Wave Height, Hs (m)", fontsize=label_font_size + 1)
ax_feet.set_ylabel("Hs (ft)", fontsize=label_font_size + 1)