Performing statistical data analysis on real-world data using the power of Python and Pandas on WoTKit sensor data.

Data Anallysis of Real-World Data With WoTKit and Pandas

#!/bin/python
import httplib, urllib, base64, json
import pandas as pd

#TODO: ADD YOUR SENSOR NAME AND CREDENTIALS
SENSOR_NAME = 'mike.yvr-arrive'
USERNAME = 'YOURWOTKITUSERNAME'
PASSWORD = 'YOUROWOTKITPASSWORD'
HOST = 'wotkit.sensetecnic.com'

auth = base64.encodestring('%s:%s' % (USERNAME, PASSWORD)).replace('\n', '')
headers = {"Authorization": "Basic %s" % auth}

def getWoTKitSensorData():
    try:
        # get 1000 datapoints
        # http://wotkit.readthedocs.org/en/latest/api_v1/api_sensor_data.html#raw-data-retrieval
        URL = "/api/v1/sensors/%s/data?beforeE=1000" % SENSOR_NAME
        conn = httplib.HTTPConnection(HOST);
        conn.request("GET", URL, headers=headers)
        data = json.loads(conn.getresponse().read())
        conn.close()
        return data
    except:
        return None

def WotKitDataToSeries(data):
    result = pd.Series()
    for i,element in enumerate(data):
        if "value" in element:
            timestamp = pd.to_datetime(element["timestamp_iso"])
            value = element["value"]
            result.set_value(timestamp, value)
    return result

if __name__ == "__main__":
    pddata = WotKitDataToSeries( getWoTKitSensorData() )
    print "Median of arrivals: %f" % pddata.median()
    print "Max arrivals: %f" % pddata.max()
    print "Variance in arrivals: %f" % pddata.var()
    print "Skewness arrivals: %f and kurtosis: %f" % (pddata.skew(), pddata.kurtosis())
    print pddata.describe()

    #Bonus: Correlation with itself, or data from another sensor:
    print "Correlation with itself: %f" % (pddata.corr(pddata))