From bd1a50d7dc3b3026090730b9853e64aa31cb9463 Mon Sep 17 00:00:00 2001
From: arokem <arokem@gmail.com>
Date: Thu, 7 May 2015 21:03:20 -0700
Subject: [PATCH] Add an example using socrata API.

---
 005-traffic-timeseries.py | 69 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 69 insertions(+)
 create mode 100644 005-traffic-timeseries.py

diff --git a/005-traffic-timeseries.py b/005-traffic-timeseries.py
new file mode 100644
index 0000000..2d45474
--- /dev/null
+++ b/005-traffic-timeseries.py
@@ -0,0 +1,69 @@
+""" 
+
+005-traffic-timeseries.py
+
+A relatively elaborate example of plotting time-series acquired through calls to the Socrata API
+
+""" 
+from datetime import datetime
+import requests
+import collections 
+import matplotlib.pyplot as plt 
+
+
+# Get traffic data:
+api_call = requests.get("https://data.seattle.gov/resource/2z5v-ecg8.json?$where=date > '2015-01-15T00:00:00'  AND date < '2015-02-15T00:00:00'")
+#store the data we just pulled down from the internet into a json object called 'all counts'
+all_counts = api_call.json()
+#create a new, empty dictionary. When we're done, this will contain counts per day
+daily_counts = {}
+#trim the time values off of data/times stamps, and add them to our empty dictionary
+for c in all_counts:
+    count_date = c['date'][:10]
+    if count_date not in daily_counts.keys():
+        daily_counts[count_date]=0
+
+#we are only interested in the data from one bike counter location, so filter out the rest
+for c in all_counts:
+    try:
+        daily_counts[c['date'][:10]] += int(c['bgt_north_of_ne_70th_total'])
+    except:
+        pass
+
+# We'll put the counts and dates into lists we will then use in plotting
+count_dates = []
+counts = []
+for key in sorted(daily_counts):
+    count_dates.append(datetime.strptime(key, '%Y-%m-%d'))
+    counts.append(daily_counts[key])
+
+
+# Repeat with calls to the API to get temperatures:
+api_call_temp = requests.get("https://data.seattle.gov/resource/egc4-d24i.json?$select=date_trunc_ymd(datetime) AS day, MAX(airtemperature) AS top_temp&$where=datetime > '2015-01-15T00:00:00' AND datetime < '2015-02-15T00:00:00' AND stationname= 'RooseveltWay_NE80thSt'&$group=day")
+raw_data = api_call_temp.json()
+daily_temps = {}
+
+for c in raw_data:
+    temp_date = c['day'][:10]
+    daily_temps[temp_date]=c['top_temp']
+
+temp_dates = []
+temps = []
+for key in sorted(daily_temps):
+    temp_dates.append(datetime.strptime(key, '%Y-%m-%d'))
+    temps.append(daily_temps[key])
+
+# We'll use the styles again
+plt.style.use('ggplot') 
+# This means: 2 rows, 1 column:
+fig, (ax1, ax2) = plt.subplots(2, 1)
+ax1.plot(temp_dates, temps)
+ax2.plot(count_dates, counts)
+plt.show()
+
+# A picture is worth 1000 words.
+
+# Challenge: plot a scatter plot showing this relationship. Consider using the 'scatter' plotting function:
+# http://matplotlib.org/examples/shapes_and_collections/scatter_demo.html
+
+# Plot a series of figures with this relationship plotted for every month in 2014
\ No newline at end of file
-- 
2.39.5