for acn, datedict in acnt_date.items(): painter.setAcntName(acn) ''' xlabel means std ''' xlabel = [] means = [] stds = [] sorted_tuple = sorted(datedict.items(), key=operator.itemgetter(0)) for tup in sorted_tuple: statis_helper.setArray(tup[1]) means.append(statis_helper.getavg()) stds.append(statis_helper.getstd()) xlabel.append(tup[0].split('_')[1]) painter.plotBar(xlabel,means,stds,acn+"_AvgTweetPerUrlAtDifferentTime") painter.plotBar(xlabel,[len(tup[1]) for tup in sorted_tuple],None,acn+"_AvgUrlAtDifferentTime") ''' acnt_url[acnt][k].append(v) ''' totalsum = [] acnt_all_url = acnt_url[acn] for v in acnt_all_url.values(): totalsum = elementAdd(totalsum,v) totalsum = elementDiv(totalsum, len(acnt_all_url)) # for painter.plotBar([i+1 for i in xrange(len(totalsum))],totalsum,None,acn+"_Url_Num")
do classification using kmeans then plot the centra ''' # from sklearn.cluster import KMeans # km = KMeans(n_clusters=3) # km.fit(avg_filt) # centras = km.cluster_centers_ xlabel = [int(x) for x in times_label] # class_percentage = {} # for i in xrange(km.n_clusters): # class_percentage[i] = 0 # for i in km.labels_: # class_percentage[i]+=1 # for i in xrange(km.n_clusters): # print class_percentage[i]*1.0/len(km.labels_) # for idx, data in enumerate(centras): # painter.plotLine(xlabel,data,str(idx)) for data in avg_filt: painter.plotLine(xlabel,data,None,'r*') painter.showplot(xlabel) #painter.showLegend() painter.newFigure() painter.plotLine(xlabel, time_counts,"number of new urls","r") def cumulativeSum(vlist,start = 0): for v in vlist: start+= v yield start painter.plotBar(xlabel,[len(t) for t in time_urls.values()],None,"Hourly URLs","Communicative URLs") painter.showplot(xlabel) print "finish!"