世界中のUFO目撃情報をk-mean法でクラスタリング

km.py

# coding:utf-8

from sklearn import cluster, datasets
import pandas as pd
import numpy as np
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

# このデータが曲者で、数値データの中に英字や記号が混ざってた
df = pd.read_csv("scrubbed.csv")
df__ = df.iloc[:,[5,9,10]].values.astype('float32')

fig = plt.figure()
ax = Axes3D(fig)

#まずは元データをプロット
#ax.scatter(df_.T[0,::1000], df_.T[1,::1000], df_.T[2,::1000])
#plt.show()

km = cluster.KMeans(n_clusters=5)
km.fit(df__)

colors = ["r","g","b","c","m"]
for i in range(300):
	ax.scatter(df__[i,0],df__[i,1],df__[i,2],c=colors[km.labels_[i]])
ax.set_xlabel(df.columns[5])
ax.set_ylabel(df.columns[9])
ax.set_zlabel(df.columns[10])
plt.show()