Snippets for Analytics

Overview

This collection contains Python snippets for data analytics. The goal is to provide a template to quickly get a preliminary overview of a new data set.

Data cleaning, overview and EDA

The Cleaning and EDA snippet contains code to perform basic checks, type conversion, data cleaning incl. imputation and outlier treatment, and exploratory data analysis (EDA).

Basic workflow

flowchart LR
subgraph 1[1. Load and Check]
    A[Load Data]--> B[Overview and Basic Checks] --> C[Type Conversion]
end
subgraph 2[2. Data Cleaning]
    C--> D[Drop NaN and Duplicates]--> E[Imputation] --> F[Treatment of Outliers]
end
subgraph 3[3. Exploratory Data Analysis - EDA]
    F--> G[Plotting univariate distributions]--> H[Plotting bivariate distributions]
end

Loading

Ensemble Models

The Ensemble methods snippet provides an example workflow to estimate different ensemble models. The snippet contains sections for 1) setting up a project, including the test-train split, 2) estimation of the models, 3) hyperparameter tuning, and 4) scoring.

Basic workflow

flowchart LR
subgraph 1[1. Setup]
    A[Imports]--> B[Test-Train Split]
end
subgraph 2[2. Estimation]
    B--> C[Bagging]
    B--> D[Gradient Boosting] 
    B--> E[Random Forest]
    B--> F[...]
end
subgraph 3[3. Tuning]
    E--> G[Grid Search]
    D--> G
    C--> G
    F--> G
end
subgraph 4[4. Scoring]
    G--> H[Accuracy, Recall, Precision, F1, ...]
end

Loading

Sklearn Pipelines

To be continued...