From: Benjamin Mako Hill <mako@atdot.cc>
Date: Sat, 9 May 2015 15:13:28 +0000 (-0700)
Subject: Merge branch 'traffic-timeseries' of github.com:arokem/matplotlib-cdsw
X-Git-Url: https://projects.mako.cc/source/matplotlib-cdsw/commitdiff_plain/6e94163bfc436e165e143cbc6259054537f554a9?hp=bd1a50d7dc3b3026090730b9853e64aa31cb9463

Merge branch 'traffic-timeseries' of github.com:arokem/matplotlib-cdsw
---

diff --git a/001-hello-plot.py b/001-hello-plot.py
new file mode 100644
index 0000000..882c22e
--- /dev/null
+++ b/001-hello-plot.py
@@ -0,0 +1,13 @@
+""" 
+
+hello_plot.py
+
+A first plot with matplotlib
+
+"""
+
+import matplotlib.pyplot as plt 
+figure, axis = plt.subplots(1)
+axis.plot([1,2,3], [2,4,8])
+plt.show()
+
diff --git a/002-subplots.py b/002-subplots.py
new file mode 100644
index 0000000..0c6749b
--- /dev/null
+++ b/002-subplots.py
@@ -0,0 +1,36 @@
+""" 
+
+Make a slightly more elaborate plot with subplots
+
+Save the figure in the end
+
+""" 
+import matplotlib.pyplot as plt
+import math
+
+# Make some data to plot
+x = []
+y1 = []
+y2 = []
+
+for i in range(100):
+    x.append(i)
+    y1.append(math.sin(i * (2 * math.pi / 100)))
+    y2.append(math.cos(i * (2 * math.pi/ 100)))
+
+# First, create an empty figure with 2 subplots
+# - The function plt.subplots returns an object for the figure and for each axes
+# - There are multiple ways to accomplish this same goal, but this is probably the
+#   simplest - notice that each subplot is associated with one of the axes objects.
+fig, (ax1, ax2) = plt.subplots(2)
+
+# Next, put one line on the first axis and both lines on the second axis
+# - On the second axes, add a legend to distinguish the two lines
+ax1.plot(x, y1)
+
+ax2.plot(x, y1, label='sin')  # The labels are what appear in the legend
+ax2.plot(x, y2, label='cos')
+ax2.legend()
+
+# Finally, save the figure as a png file
+fig.savefig('myfig.png')
\ No newline at end of file
diff --git a/003-plot-timeseries.py b/003-plot-timeseries.py
new file mode 100644
index 0000000..f0fe255
--- /dev/null
+++ b/003-plot-timeseries.py
@@ -0,0 +1,45 @@
+""" 
+003-plot-timeseries.py 
+
+Plot data from the Harry Potter data-set as a time-series
+
+""" 
+
+
+import matplotlib.pyplot as plt 
+import load_hp_data as hp
+
+# We can play with styles:
+#plt.style.use('bmh')
+plt.style.use('ggplot') 
+# To see available styles, type: 
+#plt.style.available
+
+fig, ax = plt.subplots(1)
+ax.plot(hp.columns['timestamp'],  hp.columns['size'])
+ax.set_xlabel('Time')
+ax.set_ylabel('Size of the edit')
+
+plt.show()
+
+
+# Challenge: Is edit size related to how long it's been since the last edit? 
+# => Plot the relationship between edit size and the time since the last edit:
+
+## Hint 1: the number of seconds between two edits is: 
+
+#delta_time1 = (hp.columns['timestamp'][1] - hp.columns['timestamp'][0]).total_seconds()
+
+## Hint 2: 
+
+# You can give `plt.plot` more arguments to control the shape/size/color 
+# of the markers used. For example, try: 
+
+# ax.plot([1,2,3], [2,4,8], '.')
+# ax.plot([1,2,3], [2,4,8], 'r.')
+
+# And see online documentation here: 
+# http://matplotlib.org/api/pyplot_summary.html
+# http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.plot
+
+
diff --git a/004-plot-histogram.py b/004-plot-histogram.py
new file mode 100644
index 0000000..b938fea
--- /dev/null
+++ b/004-plot-histogram.py
@@ -0,0 +1,24 @@
+"""
+004-plot-histogram.py 
+
+Plot a histogram of edit sizes 
+
+"""
+
+import matplotlib.pyplot as plt 
+import load_hp_data as hp
+
+plt.style.use('ggplot')
+
+fig, ax = plt.subplots(1)
+ax.hist(hp.columns['size'], bins=1000)
+ax.set_xlabel('Size of the edit')
+ax.set_ylabel('')
+ax.set_title('Edit size distribution')
+
+# Maybe don't really need that axis to be so long:
+# ax.set_xlim([0, 200000])
+plt.show()
+
+## Challenge : A 'mega-user' is a user with more than 1000 edits. 
+# Plot a bar chart with the maximal edit size for each one of the mega-users
diff --git a/README b/README
index 4625944..34835e1 100644
--- a/README
+++ b/README
@@ -1,4 +1,3 @@
 Using this dataset requires that you first download the following dataset:
 
-http://communitydata.cc/~mako/wikipedia_bios.csv
-
+http://communitydata.cc/~mako/hp_wiki.tsv
diff --git a/load_hp_data.py b/load_hp_data.py
new file mode 100644
index 0000000..55d497c
--- /dev/null
+++ b/load_hp_data.py
@@ -0,0 +1,31 @@
+""" load_hp_data.py 
+
+A module for loading data from the Harry Potter wikipedia data set
+
+""" 
+import csv
+from datetime import datetime
+
+f = open('hp_wiki.tsv', 'r')
+reader = csv.DictReader(f, delimiter='\t')
+
+columns = {}
+for fieldname in reader.fieldnames:
+	columns[fieldname] = []
+
+
+rows = []
+for row in reader:
+    # Convert timestamp from a string to a date:
+    row['timestamp'] = datetime.strptime(row['timestamp'], '%Y-%m-%d %H:%M:%S')
+    # Convert size from a string to an integer:
+    row['size'] = int(row['size'])
+    rows.append(row)
+
+# Sort these things, so that they give you nice ordered time-series
+sort_rows = sorted(rows, key=lambda row: row['timestamp'], reverse=False)
+
+rows = sort_rows
+for row in sort_rows:
+    for fieldname, value in row.items():
+        columns[fieldname].append(value)