f23/Cole_Lecture_Notes/39_Plotting4/Lec39_Plotting4_Solution_Nelson.ipynb

+%% Cell type:code id: tags:
+
+``` python
+# Warmup 0
+import pandas as pd
+import math
+import copy
+import requests
+import matplotlib
+import numpy as np
+from matplotlib import pyplot as plt
+matplotlib.rcParams["font.size"] = 16
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Let's compare populations from around the world!
+populations = pd.Series({
+        "China":1439323776,
+        "India": 1380004385,
+        "Mexico": 128932753,
+        "Senegal":16743927,
+        "Bahrain":1701575,
+        "Grenada":112523,
+        "Tuvalu": 11792
+})
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Plot this as a bar chart... What's the issue? It's hard to compare!
+populations.plot.bar()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Now plot it using a logarithmic scale!
+populations.plot.bar(logy = True)
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Plot the populations again but using a horizontal bar chart...
+populations.plot.barh(logx = True)
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Save the hydrants data into a variable called "hydrants"
+hydrants = pd.read_csv("Fire_Hydrants.csv")
+hydrants.head()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# What are the styles for each hydrant? Drop any NA values and
+# save to a variable called hydrant_styles
+hydrant_styles = hydrants["Style"].dropna()
+hydrant_styles
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# How many of each type are there?
+hydrant_styles.value_counts()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Clean the data first by making it uppercase...
+hydrant_styles = hydrant_styles.str.upper()
+hydrant_styles.value_counts()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# ... and just the first word
+hydrant_styles = hydrant_styles.str.split(" ").apply(lambda hyd : hyd[0])
+hydrant_styles.value_counts()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Save this to a new column called "Clean_Style"
+hydrants["Clean_Style"] = hydrant_styles
+hydrants
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Save the year_manufactured for "PACER" hydrants to a variable called pacer_years
+pacer_years = hydrants[hydrants["Clean_Style"] == "PACER"]["year_manufactured"]
+pacer_years
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Save the year_manufactured for "MUELLER" hydrants to a variable called mueller_years
+mueller_years = hydrants[hydrants["Clean_Style"] == "MUELLER"]["year_manufactured"]
+mueller_years
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Save the year_manufactured for "M-3" hydrants to a variable called m3_years
+m3_years = hydrants[hydrants["Clean_Style"] == "M-3"]["year_manufactured"]
+m3_years
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Save the year_manufactured for all other hydrants to a variable called other_years
+other_years = hydrants["year_manufactured"]
+other_years = other_years[hydrants["Clean_Style"] != "PACER"]
+other_years = other_years[hydrants["Clean_Style"] != "MUELLER"]
+other_years = other_years[hydrants["Clean_Style"] != "M-3"]
+other_years
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Drop the NA values and cast as int
+pacer_years = pacer_years.dropna().astype(int)
+mueller_years = mueller_years.dropna().astype(int)
+m3_years = m3_years.dropna().astype(int)
+other_years = other_years.dropna().astype(int)
+other_years
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Group by decade
+pacer_decades = (pacer_years // 10 * 10).value_counts()
+mueller_decades = (mueller_years // 10 * 10).value_counts()
+m3_decades = (m3_years // 10 * 10).value_counts()
+other_decades = (other_years // 10 * 10).value_counts()
+other_decades
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Make a dataframe out of this data.
+style_df = pd.DataFrame({
+    "PACER": pacer_decades.sort_index(),
+    "MUELLER": mueller_decades.sort_index(),
+    "M-3": m3_decades.sort_index(),
+    "OTHER": other_decades.sort_index()
+})
+style_df
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Visualize with a line plot.
+ax = style_df.plot.line()
+ax.set_xlabel("Decade")
+ax.set_ylabel("Hydrant Count")
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Replace the NA values with 0
+style_df = style_df.fillna(0)
+style_df.plot.line()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Visualize since only 1950 as a bar plot
+ax = style_df[style_df.index >= 1950].plot.bar()
+ax.set_xlabel("Decade")
+ax.set_ylabel("Hydrant Count")
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Stack the bar plots rather than making them appear side-by-side.
+ax = style_df[style_df.index >= 1950].plot.bar(stacked = True)
+ax.set_xlabel("Decade")
+ax.set_ylabel("Hydrant Count")
+None
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Let's put all of the above together!
+hydrants = pd.read_csv("Fire_Hydrants.csv")
+hydrants["Clean_Style"] = hydrants["Style"].dropna() \
+                            .str.upper() \
+                            .str.split(" ") \
+                            .apply(lambda hyd : hyd[0])
+
+typs = ["PACER", "MUELLER", "M-3"]
+typ_to_decades_dict = {}
+for typ in typs:
+    years = hydrants[hydrants["Clean_Style"] == typ]["year_manufactured"]
+    clean_years = years.dropna().astype(int)
+    decades = (clean_years // 10 * 10).value_counts()
+    typ_to_decades_dict[typ] = decades.sort_index()
+
+other_years = hydrants["year_manufactured"]
+for typ in typs:
+    other_years = other_years[hydrants["Clean_Style"] != typ]
+typ_to_decades_dict["OTHER"] = (other_years.dropna().astype(int) // 10 * 10).value_counts().sort_index()
+
+style_df = pd.DataFrame(typ_to_decades_dict).fillna(0)
+style_df.plot.line()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Let's remind ourselves of what the "features" of our data are.
+hydrants.columns
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Plot all of the hydrants' X and Y coordinates on a scatter plot!
+ax = hydrants.plot.scatter(x="X", y="Y", xlabel="Longitude", ylabel="Latitude")
+ax.grid()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# What about just "PACER" fire hydrants?
+pacer_hyds = hydrants[hydrants["Clean_Style"] == "PACER"]
+pacer_hyds.plot.scatter(x="X", y="Y")
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# What about just "MEDALLION" fire hydrants?
+medallion_hyds = hydrants[hydrants["Clean_Style"] == "MEDALLION"]
+ax = medallion_hyds.plot.scatter(x="X", y="Y")
+
+# What's deceiving about this? They are not on the same scale.
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# What are the minimum and maximum coordinates?
+min_x = hydrants["X"].min()
+max_x = hydrants["X"].max()
+min_y = hydrants["Y"].min()
+max_y = hydrants["Y"].max()
+
+print(min_x, max_x)
+print(min_y, max_y)
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Does the plot show all these coordinates?
+assert ax.get_xlim()[0] <= min_x
+assert ax.get_xlim()[1] >= max_x
+assert ax.get_ylim()[0] <= min_y
+assert ax.get_ylim()[1] <= max_y
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Now plot the "MEDALLION" fire hydrants using these limits...
+ax = medallion_hyds.plot.scatter(x="X", y="Y", xlim=[min_x, max_x], ylim=[min_y, max_y])
+assert ax.get_xlim()[0] <= min_x
+assert ax.get_xlim()[1] >= max_x
+assert ax.get_ylim()[0] <= min_y
+assert ax.get_ylim()[1] <= max_y
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# What if we wanted to plot 4 different types of hydrants
+# in four different "quadrants" of the graph?
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Let's dive into subplots!
+plt.subplots()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Can have a number of columns...
+plt.subplots(ncols = 2)
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# ... or rows
+plt.subplots(nrows = 2)
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Let's create some data to plot...
+s1 = pd.Series([1, 2, 3, 3, 4])
+s2 = pd.Series([5, 7, 7, 8])
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Creating the plots...
+fig, axes = plt.subplots(ncols = 2)
+
+# Plotting the data...
+s1.plot.line(ax = axes[0])
+s2.plot.line(ax = axes[1])
+# axes[0] is the area on the left, axes[1] is the area on the right
+
+# What's misleading about this? Different y axes.
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Fix this by adding sharey=True
+fig, axes = plt.subplots(ncols = 2, sharey = True)
+pd.Series([1, 2, 3, 3, 4]).plot.line(ax = axes[0])
+pd.Series([5, 7, 7, 8]).plot.line(ax = axes[1])
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# An example... Evolution of fire hydrants through the ages
+fig, axes = plt.subplots(ncols = 2, nrows = 3, sharex=True, sharey=True, figsize=(8.5, 11))
+
+hydrants[hydrants["year_manufactured"] < 1930].plot.scatter(x="X", y="Y", ax=axes[0][0], title="1920s")
+hydrants[hydrants["year_manufactured"] < 1950].plot.scatter(x="X", y="Y", ax=axes[0][1], title="1940s")
+hydrants[hydrants["year_manufactured"] < 1970].plot.scatter(x="X", y="Y", ax=axes[1][0], title="1960s")
+hydrants[hydrants["year_manufactured"] < 1990].plot.scatter(x="X", y="Y", ax=axes[1][1], title="1980s")
+hydrants[hydrants["year_manufactured"] < 2010].plot.scatter(x="X", y="Y", ax=axes[2][0], title="2000s")
+hydrants.plot.scatter(x="X", y="Y", ax=axes[2][1], title="Today")
+fig.suptitle("Evolution of Hydrants from 1920s-Today")
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Compare the locations of PACER, MUELLER, and W-59 hydrants
+# to that of the whole City of Madison
+fig, axes = plt.subplots(ncols = 2, nrows = 2, sharex=True, sharey=True)
+
+hydrants.plot.scatter(x="X", y="Y", ax=axes[0][0], title="All Hydrants")
+hydrants[hydrants["Clean_Style"] == "PACER"].plot.scatter(x="X", y="Y", ax=axes[1][0], title="Pacer", color="salmon")
+hydrants[hydrants["Clean_Style"] == "MUELLER"].plot.scatter(x="X", y="Y", ax=axes[1][1], title="Mueller", color="goldenrod")
+hydrants[hydrants["Clean_Style"] == "W-59"].plot.scatter(x="X", y="Y", ax=axes[0][1], title="W-59", color="green")
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Let's remind ourselves of the hydrant colors...
+hydrants["nozzle_color"].str.upper().value_counts()
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# What are the locations of the hydrants with
+# BLUE, GREEN, and ORANGE "nozzle_color"?
+fig, axes = plt.subplots(ncols = 3, figsize=(10, 2.5), sharex=True, sharey=True)
+hydrants[hydrants["nozzle_color"].str.upper() == "BLUE"].plot.scatter(x="X", y="Y", color="blue", ax=axes[0])
+hydrants[hydrants["nozzle_color"].str.upper() == "GREEN"].plot.scatter(x="X", y="Y", color="green", ax=axes[1])
+hydrants[hydrants["nozzle_color"].str.upper() == "ORANGE"].plot.scatter(x="X", y="Y", color="orange", ax=axes[2])
+```
+
+%% Cell type:code id: tags:
+
+``` python
+# Can you do it with less copying and pasting? Use a loop!
+colors = ["BLUE", "GREEN", "ORANGE"]
+fig, axes = plt.subplots(ncols = 3, figsize=(10, 2.5), sharex=True, sharey=True)
+
+for i in range(len(colors)):
+    colour = colors[i]
+    hyds_of_color = hydrants[hydrants["nozzle_color"].str.upper() == colour]
+    hyds_of_color.plot.scatter(x="X", y="Y", color=colour, ax=axes[i], title=colour.lower())
+```
+
+%% Cell type:code id: tags:
+
+``` python
+```