Icon

HR Use Case - Starting Point

Data Access The data of 14,000 employees are loaded. About 10 variables, partly numeric, partly categorical, Target variable: "LEFT" (0/1) indicates whether employee has resigned (1) or not (0).
Data Preparation Preparation steps, see nodes
Summary Statistics Univariate, descriptive analyses of the variables
Visualisation - Individual variables Univariate, descriptive analyses of the variables
Correlation analysis Analysis of the statistical correlation of the data
Data exploration
Statistical testing
Analytics & Modelling: Supervised machine learning 1) Forecasting / Classification Models
Data Mining Models - FOKUS: Predictive analytics Model training, testing, classification models
Logistic Regression
Decision Tree
( ... ) u.a.: (Random Forests) (Ensemble Models) (Bayes Classifier) (Linear Diskriminant-Analysis)
AutoML / Driverless AI
( ... )
Hypothesis testing We receive the data of 1000 new employees: are they comparable with our basic data set?
Analytics & Modelling: Unsupervised machine learning 2) Segmentations
Data mining models - FOKUS: structure recognition / segmentation e.g. cluster analysis
Deployment
Use of the model results in systems
( ... ) Gütekriterien u.a. (CrossValidation) (AUC) (balancing classes) (Confusion Matrix) (ROC) (R^2)
Visualisation - Several variables to each other Multivariate, descriptive analyses of the variables
Parallel Coordinates Plot (JavaScript) (legacy)
Define numeric variable as string
Number to String
Skip the not needed "ID"
Column Filter
Quitting: Red Not quitting: green
Color Manager
Row Filter (deprecated)
Statistics
Parallel Coordinates Plot (JavaScript) (legacy)
Rank Correlation
Excel Reader
Bar Chart (JavaScript) (legacy)
Data Explorer
Row Filter (deprecated)
Linear Correlation
Excel Reader
Scatter Plot (JavaScript) (legacy)
3D Scatter Plot (Plotly)
Histogram (JavaScript) (legacy)
Violin Plot (Plotly)

Nodes

Extensions

Links