Icon

Beginners - Discover the KNIME Possibilities

<p><strong>KNIME - "Discover the possibilities" (for beginners)</strong> - visual data analysis and processing</p><p>A typical KNIME workflow moves from left to right. It loads data, performs analyses and manipulations, and then exports the results for further use.<br>At any point, you can view the results and "branch off" to attach a new task to those results.<br>KNIME workflows combine functionality and documentation.</p><ul><li><p>Basic functions in KNIME and flow variables</p></li><li><p>Simple machine learning examples</p></li><li><p>Loops for automation</p></li></ul><p></p><p>Start to Learn KNIME</p><p>https://medium.com/p/d4b16c9616ad</p>

URL: KNIME Beispiel Titanic (deutsch) https://hub.knime.com/mlauber71/spaces/Public/latest/knime_einfuehrung_beispiel_titanic~QpZwCmt6Ivh_NMRT/
URL: KNIME, Paths and Loops — Automate Everything https://medium.com/p/213455788685
URL: KNIME Snippets — Rules and Conditions do apply https://medium.com/p/72b6ffbd4c9f
URL: KNIME — Machine Learning and Artificial Intelligence — A Collection https://medium.com/p/12e0f7d83b50
URL: Mastering KNIME: Unlocking Peak Performance with Expert Tips and Smart Settings https://medium.com/low-code-for-advanced-data-science/mastering-knime-unlocking-peak-performance-with-expert-tips-and-smart-settings-49e2c0dab9cc
URL: Start to Learn KNIME https://medium.com/p/d4b16c9616ad

1 - Load data - the data is located in the folder /data/ directly under the workflow. However, it could also be located elsewhere.

2 - Manipulate and merge data

3 - Export data and results

KNIME - "Discover the possibilities" (for beginners) - visual data analysis and processing

https://hub.knime.com/s/Bssy2ppqljGis_jo

A typical KNIME workflow moves from left to right. It loads data, performs analyses and manipulations, and then exports the results for further use.
At any point, you can view the results and "branch off" to attach a new task to those results.
You can start 'small' and gradually work your way forward. Old solutions continue to function smoothly.

KNIME workflows combine functionality and documentation in one!

2b - Which educational degree (education) leads to the highest income?

2a - Create graphs "How does age affect income?"

4 - Use flow variables (red lines) to dynamically filter part of the data

KNIME - Machine Learning Example

From a simple decision tree to fully automated machine learning with the most advanced platforms (H2O.ai) - all within the same interface

1 - Load data for Machine Learning (ML)

2 - Calculate ML model on the training data

3 - Save ML model

4 - Load and apply ML model on the test data

5 - ML Modell bewerten

1 - Load data

KNIME - Use loops!!

Loops are one of the most powerful tools for quickly creating dynamic automations.

3 - Export data in a loop

5 - List and reload data

Traffic light logic at each node

  • Red = Configuration not yet completed

  • Yellow = Node configured, can be executed

  • Green = Node has been executed. The results are (internally) saved and can be viewed/used

MOJO / H2O.ai models can also be used on other platforms such as R, Python, or Big Data systems

Modify individual cells in Excel based on input from KNIME

KNIME, Paths and Loops — Automate Everything

https://medium.com/p/213455788685


KNIME Snippets — Rules and Conditions do apply

https://medium.com/p/72b6ffbd4c9f

KNIME — Machine Learning and Artificial Intelligence — A Collection
https://medium.com/p/12e0f7d83b50

Decision Tree with all its exports

https://hub.knime.com/s/GbK8kmsRLtwvSKl9

KNIME and Excel — An Overview of What You Can Do

https://medium.com/p/c6d400319745

Exploring the Power of Python Graphics with KNIME:
A Collection of Examples

https://medium.com/p/841df87b5563

list_education.xlsxAn additional listthat adds text descriptions to thenumerical variable education_num (education)
Excel Reader
census_income_training.parquet=> completely unknown data
Parquet Reader
Apply model to entirely new data
Decision Tree Predictor
H2O.ai requires its own environmentin the background,which is set up here
H2O Local Context
Pass data to theH2O.ai context
Table to H2O
Automatically calculate more complex ML models here in just 60 seconds!! Adjust if necessary
H2O AutoML Learner
Age distribution by target variable (more or less than 50,000 dollars annual income)
Conditional Box Plot (JavaScript)
model_h2o_automl.zip
H2O MOJO Writer
Group byeducation
GroupBy
model_h2o_automl.zip
H2O MOJO Reader
H2O Model to MOJO
Sort in descending orderbyProportion_more_than_50k_dollars
Sorter
Rank_Income_Education
Counter Generation
Reference the text of the variable"education"by "education_num"
Joiner
Proportion_more_than_50k_dollars
Math Formula
$education$ = $${Seducation}$$ => TRUE
Rule-based Row Filter
Create Path Variables
Create File/Folder Variables
Binary Classification Inspector
$Counter$ = 1 => TRUE
Rule-based Row Filter
Display the variable: "Education" with the highestaverage income
Table Row to Variable
train.table
Table Reader
"Target" Variableas String
Number to String
Keep only the columns for machine learning
Column Filter
Create a simplemachine learning model
Decision Tree Learner
Apply model to 30% test data
Decision Tree Predictor
70% / 30%Separationfor Training and Test
Partitioning
View the decision tree
Decision Tree View (JavaScript)
Binary Classification Inspector
Domain Calculator
model_decision_tree.pmml
PMML Reader
Output of the selectionEducation with the highest average incomeas an Excel file
Excel Writer
model_decision_tree.pmml
PMML Writer
Convert the tree model into SQL
PMML to SQL
Replace missing values with-999 or _NA_
Missing Value
Educational degree by income.xlsx
Excel Writer
Corrected Excel CellsB1 - Number of respondentsC1 - Respondents earning over 50k dollars
Table Creator
LOOP END Import
Loop End
Model
Constant Value Column
Import each subgroup as an Excel file Subgroup_*.xlsx
Excel Reader
LOOP START Import
Table Row to Variable Loop Start
Image to Table
Educational degree by income - corrected.xlsx
Excel Cell Updater
Table to Image
age_vs_income.png
Image Writer (Port)
Model
Constant Value Column
2.560x1.440 PNG file
Renderer to Image
Compare H2O.ai andDecision Tree
Concatenate
LOOP END
Variable Loop End
List existing files Subgroup_*.xlsx
List Files/Folders
Create path variable with a dynamic name
Create File/Folder Variables
Apply model to 30% test data
H2O MOJO Predictor (Classification)
martial_status
GroupBy
$Count(Target)$ >= 100 => TRUEat least 100 cases in the group
Rule-based Row Filter
Get a quick overviewof the data
Data Explorer
train.table
Table Reader
var_file_name
Java Edit Variable (simple)
Output each subgroup as an Excel file Subgroup_*.xlsx
Excel Writer
LOOP START
Table Row to Variable Loop Start
$marital_status$ = $${Smarital_status}$$ => TRUE
Rule-based Row Filter
train.table=> used later for Machine Learning
Table Writer
UCI "Census Income Data Set"Preparation => a metanode encapsulates various functions
initial Data preparation
Binary Classification Inspector
census_income_training.parquetDemographic data from the USA1996 - whether someone earns more than$50,000 per year.Also known as the "Adult" datasethttps://archive.ics.uci.edu/ml/datasets/census+income
Parquet Reader
if (column("income_50k") == ">50K") {1}else if (column("income_50k") == "<=50K") {0}else {null} Create a Variable "Target" 0/1
Column Expressions

Nodes

Extensions

Links