Py Tip: Siuba

Author: mdancho84

.vscode/settings.json

@@ -1,5 +1,5 @@
 {
-    "python.pythonPath": "C:\\Users\\mdanc\\Anaconda3\\envs\\ds4b_101p\\python.exe",
+    "python.pythonPath": "/Users/mdancho/opt/anaconda3/envs/free_python_tips/bin/python",
     "jupyter.sendSelectionToInteractiveWindow": true,
     "jupyter.notebookFileRoot": "${workspaceFolder}",
 }

02_siuba/02_siuba.py

@@ -0,0 +1,60 @@
+# PYTHON TIPS ----
+# TIP 002 | Siuba: Dplyr for Python ----
+#
+# 👉 For Weekly Python-Tips, Sign Up Here:
+#    https://mailchi.mp/business-science/python_tips_newsletter
+
+# LIBRARIES ----
+import numpy as np
+import pandas as pd
+
+from siuba import _
+from siuba.dply.verbs import group_by, mutate, select, summarize, ungroup
+
+# DATASET ----
+
+mpg_df = pd.read_csv("https://raw.githubusercontent.com/mwaskom/seaborn-data/master/mpg.csv")
+mpg_df
+
+# 1.0 GROUP BY + SUMMARIZE
+# Goal: Mean and Standard Deviation of weight by engine size
+
+weight_by_cyl_df = mpg_df >> \
+    group_by("cylinders") >> \
+    summarize(
+        mean_weight = np.mean(_.weight),
+        sd_weight   = np.std(_.weight)
+    )
+
+weight_by_cyl_df
+
+# 2.0 GROUP BY + MUTATE
+# Goal: De-mean the mpg by average of each cylinder
+
+mpg_demeaned_by_cyl_df = mpg_df >> \
+    select('name', 'cylinders', 'mpg') >> \
+    group_by("cylinders") >> \
+    mutate(
+        mean_mpg = np.mean(_.mpg)
+    ) >> \
+    ungroup() >> \
+    mutate(
+        mpg_demeaned_by_cyl = _.mpg - _.mean_mpg
+    )
+
+mpg_demeaned_by_cyl_df
+
+# 3.0 PANDAS 
+mpg_demeaned_by_cyl_df[['name', 'cylinders', 'mpg_demeaned_by_cyl']] \
+    .sort_values('mpg_demeaned_by_cyl', ascending = False) \
+    .style \
+    .background_gradient()
+
+# LEARNING PANDAS ----
+# - Siuba is great for when you are coming from R to Python (like me)
+# - Teams use Pandas: 99% of data wranlging code is written with Pandas
+# - Better Learn Pandas if you want to be part of the Team
+
+# I TEACH PANDAS (FROM AN R-USERS PERSPECTIVE)!
+# Python for Data Science Automation Course (Contains 5 hours of Pandas)
+# https://university.business-science.io/p/python-for-data-science-automation-ds4b-101p

environment.yml

@@ -0,0 +1,90 @@
+# In terminal, run: conda env create -f environment.yml
+# To update, run: conda env update -f environment.yml
+name: free_python_tips
+channels:
+  - anaconda
+  - conda-forge
+  - defaults
+dependencies:
+  - python=3.7.1
+  - pip
+  - pip:
+      # Core Data
+      - numpy==1.20.2
+      - pandas==1.2.2
+
+      # R Data
+      - plydata==0.4.3
+      - siuba==0.0.24
+      - datatable
+
+      # Visualization
+      - matplotlib==3.3.4
+      - plotnine==0.7.1
+      - mizani==0.7.2
+      - plotly==4.14.3
+      - altair==4.1.0
+
+      # EDA
+      - pandas-profiling
+      - ppscore==1.2.0
+      - pyjanitor==0.20.14
+
+      # Modeling & Machine Learning
+      - statsmodels
+      - nltk==3.5
+      - h2o==3.32.0.3
+      - pycaret==2.3.0
+      - scikit-learn==0.23.2
+      - xgboost==0.90
+      - lightgbm==3.1.1
+      - catboost==0.24.4
+      - sklearn-pandas==2.0.4
+      - scikit-misc==0.1.3
+
+      # Time Series
+      - sktime==0.5.3
+      - pmdarima==1.8.1
+      - tsfresh==0.17.0
+
+      # Scalability & Automation
+      - dask==2.30.0
+      - dask-ml==1.8.0
+      - dask-xgboost==0.1.11
+      - zict==1.0.0
+      - joblib==1.0.1
+
+      # API
+      - fastapi==0.63.0
+      - uvicorn==0.13.4
+
+      # Database
+      - sqlalchemy==1.4.7
+
+      # Excel
+      - xlsxwriter==1.3.7
+      - openpyxl
+
+      # Jupyter
+      - jupyterlab==3.0.13
+      - jupyterlab-server==2.4.0
+      - jupyter-packaging==0.7.12
+      - jupyter-server==1.6.1 # Solves ImportError: cannot import name 'get_version_info' from 'jupyter_packaging'
+      - ipywidgets==7.6.3
+      - ipympl==0.7.0
+      - jupytext
+      - papermill==2.3.3
+      - nbconvert==5.6.1
+
+      # Apps
+      - streamlit==0.80.0
+
+      # Terminal Formatting
+      - rich
+
+      # Extending Pandas
+      - pandas_flavor
+
+      # R users
+      - radian
+      - jedi==0.17.2