Installing orbit with pip on Windows 10

[dishkakrauch]

Describe the bug Orbit can be installed with pip but it cannot be fitted with pystan backend in case of not supported MSVC compiler.

To Reproduce

1. Install Visual Studio 2015/2017/2019 with default C++ compilers including MSVC.
2. Install pystan with conda: conda install pystan
3. Install orbit with pip: pip install orbit-ml

Expected behavior This code should go without any error:

import matplotlib.pyplot as plt

import orbit
from orbit.models.dlt import DLTMAP, DLTFull
from orbit.diagnostics.plot import plot_predicted_data, plot_predicted_components
from orbit.utils.dataset import load_iclaims
from orbit.utils.plot import get_orbit_style

# load log-transformed data
df = load_iclaims()
train_df = df[df['week'] < '2017-01-01']
test_df = df[df['week'] >= '2017-01-01']

response_col = 'claims'
date_col = 'week'
regressor_col = ['trend.unemploy', 'trend.filling', 'trend.job']

dlt = DLTFull(
    response_col=response_col,
    regressor_col=regressor_col,
    date_col=date_col,
    seasonality=52,
    prediction_percentiles=[5, 95],
)

dlt.fit(train_df)

Screenshots

Environment (please complete the following information):

• OS: Windows 10
• Python Version: 3.8.5
• Versions of Major Dependencies (pandas, scikit-learn, cython): [e.g. pandas==1.1.3, scikit-learn==0.22.1, cython==0.29.21]

Additional context Prophet works well with this setup.

[wangzhishi]

hi @dishkakrauch , this seems to be an issue related to installation pystan 2 on windows. I recommend to follow their guidance to install pystan on windows. After successfully installing pystan, then install orbit and try it out.

Pystan is a requirement for orbit, and it is tricky to use pystan on windows OS. The issue you raised is one problem one level lower before orbit.