def get_barometer(API_KEY,myPostCode,sensor_locations):
opensensor = London(API_KEY)
weight = opensensor.get_weights(myPostCode,sensor_locations)
#the first date we try and get data
start_date = (datetime.datetime.now() + datetime.timedelta(hours=-8)).strftime("%Y-%m-%dT%H:00:00.000Z")
#the last date defaults to today
end_date = (datetime.datetime.now() + datetime.timedelta(days=1)).strftime("%Y-%m-%d")
NO2 = pandas.DataFrame()
for row in weight.index:
mtx = opensensor.getAirQuality(start=start_date,end=end_date,location = row).get('NO2')
if mtx is not None:
NO2[row] = mtx[row]
#grab the NO2 data where we bhave the intersect of weights and NO2
weight = weight[list(NO2.columns)]/weight[list(NO2.columns)].sum()
#sort the matrices
NO2 = NO2.reindex_axis(sorted(NO2.columns), axis=1)
weight = weight.reindex(sorted(weight.index))
#ffill and bfill NA's
NO2 = NO2.ffill().bfill()
#the AQ vector for my post code
myAQ = (NO2 * weight).sum(axis = 1)
#set the decay parameters for the exp moving avgs
nslow = 6
nfast = 3
#get the MACD and exp moving avgs
NO2_emafast,NO2_emaslow,NO2_macd = opensensor.getMovingAverageConvergence(myAQ, nslow=nslow, nfast=nfast)
#make into 1/0 signal and get the last reading
signal = (-NO2_macd/abs(NO2_macd))[len(NO2_macd) - 1]
return signal
opensensor = London(API_KEY)
#Get the meta data not on opensensors at the moment
with open(LAQN_file) as f:
reader = csv.reader(f)
LAQN_locations = dict((rows[0],[float(rows[1]),float(rows[2])]) for rows in reader)
def moving_average_convergence(x, nslow=8, nfast=4):
"""
wrap the method in the London object so we can use apply
"""
NO2_emafast,NO2_emaslow,NO2_macd = opensensor.getMovingAverageConvergence(x, nslow, nfast)
return pandas.Series(NO2_macd)
#get the list of LAQN measures
LAQN_mtx = opensensor.getAirQuality(start=start_date,end=end_date)
#get bike data
bikes_mtx, bikes_metadata = opensensor.getTFLBikes((datetime.datetime.now() + datetime.timedelta(days=-1)).strftime("%Y%m%d"),end_date)
coverage_day = LAQN_mtx['NO2'].transpose().apply(np.isnan).sum()/LAQN_mtx['NO2'].transpose().apply(np.isnan).count()
NO2 = LAQN_mtx['NO2'][coverage_day < 0.2]
coverage_sensor = NO2.apply(np.isnan).sum()/NO2.apply(np.isnan).count()
NO2 = NO2.transpose()[coverage_sensor < 0.2].transpose()
mean_daily_variance = NO2.groupby(lambda x: (x.date)).std()
for row in mean_daily_variance.index: