def graph2(Data , Station, Constituent ):
Unit= ""
Unit = units(Constituent )
showflier = False #outliers
Data= Data.dropna(subset = [Constituent])
DataX = data(Data , Station)
if DataX[Constituent].sum() == 0:
import pandas as pd
return pd.DataFrame({'A' : []}) , pd.DataFrame({'A' : []})
else:
DataX.boxplot(column = [Constituent],by='months2' , figsize = (15,6),showfliers=showflier)
plt.ylabel(Constituent+Unit[0])
plt.xlabel('Months, n = ' + str(DataX[Constituent].count()))
plt.suptitle("")
plt.title('Boxplot of Ambient '+Constituent+' over months' + ' ('+Station+')')
plt.gcf().autofmt_xdate()
if Unit[1]!=0:
plt.axhline(y=Unit[1], color='r', linestyle=':')
if Unit[2]!=0:
plt.axhline(y=Unit[2], color='r', linestyle=':')
plt.show()
stats = DataX [[ Constituent , 'months2']].groupby('months2').describe()
stats2 = DataX [[ Constituent ]].describe()
Violation = DataX[DataX[Constituent].gt(Unit[2])]
#stats3 = Violation[[ Constituent , 'months2']].groupby('months2').describe()
#stats3['% Exceedence'] = stats3[stats3.columns[0]]/stats[stats.columns[0]]*100
return stats.round(decimals=2) , stats2.round(decimals=2) #, stats3.round(decimals=2)
def AnaVSPot2( DataX , Constituent , Station ):
import seaborn as sns
Data= DataX.dropna(subset = [Constituent])
DataX=Data
#error if there is no data
if DataX[Constituent].sum() == 0:
return "No data here"
else:
DataX['stamp'] = (DataX['months2']-0.5)/12 + DataX['year']
Unit = ""
Unit = units(Constituent)
plt.rcParams.update({'font.size': 16})
sns.lmplot( x="stamp", y=Constituent, data=DataX, fit_reg=False, hue='Watershed' , size = 7 , aspect = 2)
plt.title('Scatterplot of Ambient '+Constituent+', Comparison between Anacostia and Potomac (all stations)')
plt.ylabel(Constituent+Unit[0])
if Unit[1]!=0:
plt.axhline(y=Unit[1], color='r', linestyle=':')
if Unit[2]!=0:
plt.axhline(y=Unit[2], color='r', linestyle=':')
plt.show()
return 'Scatterplot of Ambient '+Constituent+', Comparison between Anacostia and Potomac (all stations)'
def AnaVSPot( DataX , Constituent , Station):
import seaborn as sns
Data= DataX.dropna(subset = [Constituent])
DataX=Data
#error if there is no data
if DataX[Constituent].sum() == 0:
return "No data here"
else:
Unit = ""
Unit = units(Constituent )
plt.figure(3, figsize = (12,12))
plt.rcParams.update({'font.size': 16})
plt.title('Boxplot of Ambient '+Constituent+', Comparison between Anacostia and Potomac (all stations)')
sns.boxplot(x="year", y= Constituent, hue="Watershed", data=DataX ,showfliers=False)
plt.ylabel(Constituent+Unit[0])
if Unit[1]!=0:
plt.axhline(y=Unit[1], color='r', linestyle=':')
if Unit[2]!=0:
plt.axhline(y=Unit[2], color='r', linestyle=':')
plt.show()
return 'Boxplot of Ambient '+Constituent+', Comparison between Anacostia and Potomac (all stations)'
def graph(Data , Station, Constituent ):
Unit= ""
Unit = units(Constituent )
Data= Data.dropna(subset = [Constituent])
DataX = data(Data , Station)
Data2 = DataX[Constituent].as_matrix() #matrix
Data2=Data2[~np.isnan(Data2)]
#error if there is no data
if Data2.sum() == 0:
return "No data here"
else:
#CDF and Histogram
plt.figure(1, figsize = (12,12))
plt.rcParams.update({'font.size': 22})
plt.subplot(211)
plt.title('Distribution of Ambient '+Constituent+' data' + ' ('+Station+')' )
plt.grid(True)
plt.hist(Data2, 50, facecolor='green', edgecolor='blue' )
plt.ylabel('Frequency')
if Unit[1]!=0:
plt.axvline(x=Unit[1], color='r', linestyle=':')
if Unit[2]!=0:
plt.axvline(x=Unit[2], color='r', linestyle=':')
plt.subplot(212)
plt.plot(np.sort(Data2), np.linspace(0,1,(Data2.size)))
plt.xlabel(Constituent+Unit[0] + ' n= ' + str(Data2.size))
plt.ylabel('Probability')
plt.grid(True)
if Unit[1]!=0:
plt.axvline(x=Unit[1], color='r', linestyle=':',linewidth=2)
if Unit[2]!=0:
plt.axvline(x=Unit[2], color='r', linestyle=':',linewidth=2)
plt.show()
plt.rcParams.update({'font.size': 12})
fifthP = np.nanpercentile(Data2,5)
tenthP = np.nanpercentile(Data2,10)
twentyfifthP = np.nanpercentile(Data2,25)
fiftiethP = np.nanpercentile(Data2,50)
SeventyFifthP = np.nanpercentile(Data2,75)
NintiethP = np.nanpercentile(Data2,90)
NintyfifthP = np.nanpercentile(Data2,95)
Average = np.average(Data2)
stats = '5th Pi = ' + str(round(fifthP,2))+'10th Pi = ' + str(round(tenthP,2)) + ', 25th Pi = ' + str(round(twentyfifthP,2)) +', Median = ' + str(round(fiftiethP,2)) +', 75th Pi = ' + str(round(SeventyFifthP,2)) +', 90th Pi = ' + str(round(NintiethP,2))+', 95th Pi = ' + str(round(NintyfifthP,2)) +', Average = ' + str(round(Average,2))
#return 'Distribution of Ambient '+Constituent+' data' + ' ('+Station+')'
return stats
def AnaVSPot4( DataX , Constituent , Station):
import seaborn as sns
Data= DataX.dropna(subset = [Constituent])
DataX=Data
#error if there is no data
if DataX[Constituent].sum() == 0:
return "No data here"
else:
if Station[0] =='All Stations':
DataX=Data
elif Station[0] in {"AAG01", "AAG02", "ANA01", "ANA02", "ANA03", "ANA04", "ANA05", "ANA06", "ANA07", "ANA08", "ANA09", "ANA10", "ANA11", "ANA12", "ANA13", "ANA14", "ANA15", "ANA16", "ANA17", "ANA18", "ANA19", "ANA20", "ANA21", "ANA21 ", "ANA22", "ANA23", "ANA24", "ANA25", "ANA26", "ANA27", "ANA29", "ANA30" ,"TDU01", "TFC01", "TFD01", "TFE01", "TFS01", "THR01", "TNA01", "TNS01", "TOR01", "TPB01", "TTX27", "TUT01", "TWB01", "TWB02", "TWB03", "TWB04", "TWB05", "TWB06" , "TFS01" ,"KNG01", "KNG02" ,"PMS01", "PMS02", "PMS03", "PMS05", "PMS07", "PMS08", "PMS09", "PMS10", "PMS11", "PMS12", "PMS13", "PMS16", "PMS18", "PMS21", "PMS21 ", "PMS23", "PMS25", "PMS27", "PMS29", "PMS31", "PMS33", "PMS35", "PMS37", "PMS39", "PMS41", "PMS44", "PMS46", "PMS48", "PMS51" , "PMS52" ,"TBK01", "TBR01", "TCO01", "TCO06" , "TDA01" , "TDO01" , "TFB01" , "TFB02","RCR01", "RCR04","RCR07", "RCR09" ,"TKV01""TLU01", "TMH01", "TPI01", "TPO01", "TPY01", "TSO01" ,"PWC04" ,"PTB01" ,"CHAIN"}:
DataX =DataX[DataX['Station'].isin(Station)]
else:
DataX =DataX[DataX['Watershed'].isin(Station)]
if DataX[Constituent].sum() == 0:
return "No data here"
Unit = ""
Unit = units(Constituent )
plt.figure(3, figsize = (12,12))
plt.rcParams.update({'font.size': 16})
plt.title('Boxplot of Ambient '+Constituent+', Comparison between Anacostia and Potomac (all stations)')
ax = sns.boxplot(x="year", y= Constituent, hue="Watershed", data=DataX ,showfliers=False)
ax.set_xticklabels(ax.get_xticklabels(),rotation=90)
plt.ylabel(Constituent+Unit[0])
plt.xlabel('Years, n = ' + str(DataX[Constituent].count()))
if Unit[1]!=0:
plt.axhline(y=Unit[1], color='r', linestyle=':')
if Unit[2]!=0:
plt.axhline(y=Unit[2], color='r', linestyle=':')
plt.show()
return 'Boxplot of Ambient '+Constituent+', Comparison between Anacostia and Potomac (all stations)'
def AnaVSPot3( DataX , Constituent , Station):
import seaborn as sns
Data= DataX.dropna(subset = [Constituent])
DataX = data(Data , Station)
if DataX[Constituent].sum() == 0:
import pandas as pd
return pd.DataFrame({'A' : []}) , pd.DataFrame({'A' : []})
else:
Unit = ""
Unit = units(Constituent )
plt.figure(3, figsize = (20,12))
plt.rcParams.update({'font.size': 16})
ax = sns.boxplot(x="year", y= Constituent, hue="quarter", data=DataX,showfliers=False )
ax.set_xticklabels(ax.get_xticklabels(),rotation=90)
plt.title('Boxplot of Ambient '+Constituent+' by year and quarter'+ ' ('+Station+')')
plt.ylabel(Constituent+Unit[0])
if Unit[1]!=0:
plt.axhline(y=Unit[1], color='r', linestyle=':')
if Unit[2]!=0:
plt.axhline(y=Unit[2], color='r', linestyle=':')
plt.show()
stats = DataX [[ Constituent , 'quarter']].groupby('quarter').describe()
stats2 = DataX [[ Constituent ]].describe()
# stats3 = DataX [[ Constituent , 'YearQuart']].groupby('YearQuart').describe()
# Violation = DataX[DataX[Constituent].gt(Unit[2])]
# stats3 = Violation[[ Constituent , 'quarter']].groupby('quarter').describe()
# stats3['% Exceedence'] = stats3[stats3.columns[0]]/stats[stats.columns[0]]*100
return stats.round(decimals=2) , stats2.round(decimals=2) # , stats3.round(decimals=2)