# coding: utf-8

# # Baseline model and assessment - Uses 100 days rather than cumsum versus unadjusted popularity

# In[20]:

import pandas as pd
import numpy as np
get_ipython().run_line_magic('matplotlib', 'inline')

import matplotlib.pyplot as plt


# In[13]:

df = pd.read_feather('../data/basemodel.feather')
df.head()


# In[48]:

from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures
from sklearn.pipeline import make_pipeline
from sklearn.grid_search import GridSearchCV
from sklearn.metrics import r2_score
from sklearn.metrics import mean_squared_error
model = LinearRegression(fit_intercept = True)
x = df[['popularity']]
y = df[['log_all_streams']]

def PolynomialRegression(degree=2, **kwargs):
    return make_pipeline(PolynomialFeatures(degree), LinearRegression(**kwargs))

param_grid = {'polynomialfeatures__degree': np.arange(3), 'linearregression__normalize': [True, False]}

grid = GridSearchCV(PolynomialRegression(), param_grid, cv = 5, error_score=mean_squared_error)
grid.fit(x,y)


# In[49]:

model = grid.best_estimator_
y_pred = model.predict(x)

plt.scatter(y_pred, y_pred - y)
plt.hlines(y = 0, xmin = 6, xmax = 10)

rmse = np.sqrt(np.sum((y_pred - y)**2))
print(f'RMSE: {rmse}')
print(f'Rsq: {r2_score(y,y_pred)}')


# In[50]:

plt.scatter(x,y)
lim = plt.axis()
plt.scatter(x,y_pred, c = 'red')


# ## Next iteration
# * Evaluate outlier at 0 popularity. It seems to skew the results. It looks like linear would be a good fit here actually

# In[52]:

df.sort_values('popularity')