用Python实现数据化运营分析实例：销售预测

案例场景：每个销售型公司都有一定的促销费用，促销费用可以带来销售量的显著提升；当给出一定的促销费用时，预计会带来多大的商品销售量？

数据源地址：https://pan.baidu.com/s/1VE7zmWiToI5A4zuc8zmZaQ

# 导入库
import re
import numpy
from sklearn import linear_model
from matplotlib import pyplot as plt

# 导入数据
fn = open('data.txt', 'r')
all_data = fn.readlines()
fn.close()

In [10]: all_data[-5:]
Out[10]:
['21511.0\t59960.0\n',
 '28166.0\t85622.0\n',
 '34130.0\t82463.0\n',
 '17789.0\t64759.0\n',
 '21382.0\t54315.0\n']

# 数据预处理
x = []
y = []
for single_data in all_data:
    tmp_data = re.split('\t|\n', single_data)
    x.append(float(tmp_data[0]))
    y.append(float(tmp_data[1]))

x = numpy.array(x).reshape([-1, 1])
y = numpy.array(y).reshape([-1, 1])

# 数据分析展示
plt.scatter(x, y)
plt.show()

# 数据建模
model = linear_model.LinearRegression()
model.fit(x, y)

# 模型评估
##针对线性模型y=ax+b，由系数a和截距b两个参数;
model_coef = model.coef_        #系数a
model_intercept = model.intercept_         #截距b
r2 = model.score(x, y)

In [23]: r2 = model.score(x,y)

In [24]: model_coef
Out[24]: array([[ 2.09463661]])

In [25]: model_intercept
Out[25]: array([ 13175.36904199])

In [26]: r2
Out[26]: 0.78764146847589545


# 销售预测
In [27]: new_x = 84610

In [28]: pre_y = model.predict(new_x)

In [29]: pre_y
Out[29]: array([[ 190402.57234225]])