jcheong0428

# Install pystan
!pip install -q pystan
# Install Facebook Prophet 
!pip install -q fbprophet
# Download Google Trends data
!git clone https://gist.github.com/jcheong0428/838a1b1daf67473373da50627bc1bbda
!cp 838a1b1daf67473373da50627bc1bbda/*.csv .

import pandas as pd
import seaborn as sns

jcheong0428

jcheong0428

jcheong0428

Category: All categories

Month,back pain: (United States),lower back pain: (United States),neck pain: (United States)
2004-01,29,8,7
2004-02,31,7,9
2004-03,33,9,7
2004-04,30,7,7
2004-05,35,8,9
2004-06,37,10,7
2004-07,33,9,9

jcheong0428

jcheong0428

# Install pymer4
!pip install -q pymer4
# load pymer4
from pymer4.models import Lmer
model = Lmer('Weight ~ Time + Evit + (1 + Time|Pig)', data=data)
display(model.fit())
# ANOVA results from fitted model
display(model.anova())
# Plot estimated model coefficients 
model.plot_summary()

jcheong0428

%%R -i data
# Specify Evit as a factor and set a polynomial contrast.
data$Evit <- as.factor(data$Evit)
contrasts(data$Evit) <- contr.poly 
m<-lmer('Weight ~ Time * Evit + (1|Pig)', data=data)
print(summary(m))

# Relevel categorical data so base level is Evit100
m<-lmer('Weight ~ Time * relevel(Evit, ref="Evit100") + (1|Pig)', data=data)
print(summary(m))

jcheong0428

%%R -i data -o betas
m<-lmer('Weight ~ Time + (1+Time|Pig)', data=data)
print(summary(m))
betas <- fixef(m)

jcheong0428

# make sure rpy2 was loaded using %load_ext rpy2.ipython
%%R
library(lme4)
library(lmerTest)
data(dietox, package='geepack')
m<-lmer('Weight ~ Time + (1+Time|Pig)', data=dietox)
print(summary(m))

jcheong0428

!apt-get install r-base
!pip install -q rpy2
packnames = ('lme4', 'lmerTest', 'emmeans', "geepack")
from rpy2.robjects.packages import importr
from rpy2.robjects.vectors import StrVector
utils = importr("utils")
utils.chooseCRANmirror(ind=1)
utils.install_packages(StrVector(packnames))