#Plotting 3D graph
q=0
r=0
s=0
fig2, axes2 = plt.subplots(nrows=7, ncols=2, sharex=True)

for b in Brokers:
    Broker_Clusters=TimeSeriesData[TimeSeriesData.BrokerName == b][['Gross_Premium','Year','Month']]

    n_cluster = range(1, 20)

    kmeans = [KMeans(n_clusters=i).fit(Broker_Clusters) for i in n_cluster]
    scores = [kmeans[i].score(Broker_Clusters) for i in range(len(kmeans))]
    km = KMeans(n_clusters=broker_KMean[j])
    km.fit(Broker_Clusters)
    labels = km.labels_

    s=s+1

    if (q % 2) == 0:
        #a=df.plot(ax=axes[n,0],y=x,label="Observed")
        #pred_sarima.plot(ax=a,y=x,linestyle='--',label="Prediction")
        axes2[r,0]=Axes3D(fig2, rect=[0, 0, 0.95, 1], elev=48, azim=134)
        axes2[r,0].scatter(Broker_Clusters.iloc[:,0], Broker_Clusters.iloc[:,1],Broker_Clusters.iloc[:,2],c=labels.astype(np.float), edgecolor="k")
        axes2[r,0].set_xlabel("Gross Premium")
        axes2[r,0].set_ylabel("Year")
        axes2[r,0].set_zlabel("Month")
        q=q+1
        r = r

    elif (q % 2) == 1:
        #a=df.plot(ax=axes[j,1],y=x,label="observed")
        #pred_sarima.plot(ax=a,y=x,linestyle='--',label="Prediction")
        axes2[r,1]=Axes3D(fig2, rect=[0, 0, 0.95, 1], elev=48, azim=134)
        axes2[r,1].scatter(Broker_Clusters.iloc[:,0], Broker_Clusters.iloc[:,1],Broker_Clusters.iloc[:,2],c=labels.astype(np.float), edgecolor="k")
        axes2[r,1].set_xlabel("Gross Premium")
        axes2[r,1].set_ylabel("Year")
        axes2[r,1].set_zlabel("Month")
        q=q+1
        r = r+1
    Broker_Clusters = Broker_Clusters.reset_index(drop=True)