v.3.7

Maintenance Release

Code improvements and bug fixes.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik. In addition, models can be interpreted using Skater.
• Unsupervised Machine Learning: Also implemented using scikit-learn. This provides capabilities for dimensionality reduction and clustering.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.7:

• Fix for the Preprocessor class in _machine_learning.py to handle multiprocessing
• Docker image update to use Python 3.6.7 and published to Docker Hub

qlik-py-tools-3.7.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.3.6

Maintenance Release

Important fixes for machine learning with larger datasets.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik. In addition, models can be interpreted using Skater.
• Unsupervised Machine Learning: Also implemented using scikit-learn. This provides capabilities for dimensionality reduction and clustering.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.6:

• Fix for feature importance failing for > 5000 samples
• Fix for Skater sending duplicate samples to sklearn pipeline
• Fix for hashing columns with null values after aligning with training data
• Added powershell scripts for installing and running the SSE
• Updated Docker image published to Docker Hub

qlik-py-tools-3.6.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.3.5

Feature Release

New functionality to perform k-fold cross validation.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik. In addition, models can be interpreted using Skater.
• Unsupervised Machine Learning: Also implemented using scikit-learn. This provides capabilities for dimensionality reduction and clustering.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.5:

• Added functionality for k-fold cross validation
• New sample apps and documentation for k-fold cross validation
• Updates for key driver analysis using Skater.

qlik-py-tools-3.5.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.3.4

Feature Release

New functionality added for interpreting models using Skater.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik. In addition, models can be interpreted using Skater.
• Unsupervised Machine Learning: Also implemented using scikit-learn. This provides capabilities for dimensionality reduction and clustering.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.4:

• Updates for key driver analysis using Skater.
• Bug fix for regression when using the predict function.
• Bug fix for feature importances when the dataset has missing values.

qlik-py-tools-3.4.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.3.3

Feature Release

New functionality added for interpreting models using Skater.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik. In addition, models can be interpreted using Skater.
• Unsupervised Machine Learning: Also implemented using scikit-learn. This provides capabilities for dimensionality reduction and clustering.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.3:

• Added capabilities for getting model agnostic feature importances using Skater.
• Updated the scikit-learn Train & Test sample app to provide feature importances.

qlik-py-tools-3.3.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.3.2

Feature Release

This release adds functionality for unsupervised machine learning using scikit-learn.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik.
• Unsupervised Machine Learning: Also implemented using scikit-learn. This provides capabilities for dimensionality reduction and clustering.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.2:

• Added capabilities for clustering and matrix decomposition/ dimensionality reduction.
• Additional feature preparation strategies: count vectorization and term frequency, inverse document frequency (TF-IDF).
• Bundling of responses to allow for larger message sizes.
• Fixed and enhanced debug outputs

qlik-py-tools-3.2.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.3.1

Minor Release

This release adds on to the predictive analytics functionality for Qlik.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.1:

• Added parameter tuning functionality for supervised machine learning.
• Added a sample app for parameter tuning.

qlik-py-tools-3.1.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.3.0

Major Release

This is a major release that introduces functions for predictive analytics in Qlik.

This release includes:

• Supervised Machine Learning : Implemented using scikit-learn, the go-to machine learning library for Python. This SSE implements the full machine learning flow from data preparation, model training and evaluation, to making predictions in Qlik.
• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis : Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.3.0:

• New functions for training, testing and evaluating predictive models and then using them for making predictions in real-time.
• Updated installation script and steps to avoid issues in Windows with Python 3.7, pystan, and C++ compilers.
• Changed how the SSE converts strings to numbers so that it is independent of the system's locale settings.

qlik-py-tools-3.0.zip

This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.2.4

Minor Release

This release includes functions for clustering, forecasting and linear correlations.

• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis: Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.2.4:

• Added a new parameter for Prophet that allows multiplicative seasonality. This is useful if the seasonality is not a constant additive factor as assumed by default in Prophet.
• Fixed a bug for the Prophet functions which gave unexpected behavior for the return=y_then_yhat parameter with take_log=true.

qlik-py-tools-2.4.zip
This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.

v.2.3

Minor Release

This release includes functions for clustering, forecasting and linear correlations.

• Clustering : Implemented using HDBSCAN, a high performance algorithm that is great for exploratory data analysis.
• Time series forecasting : Implemented using Facebook Prophet, a modern library for easily generating good quality forecasts.
• Seasonality and holiday analysis: Also using Facebook Prophet.
• Linear correlations : Implemented using Pandas.

Change Log v.2.3:

• The SSE now uses locale settings to get the number format when converting strings to numbers. This approach was taken for Clustering and has now been applied to the Prophet class as well.

qlik-py-tools-2.3.zip
This zip archive only contains the files needed to deploy the SSE. To get the sample apps download the full source code above or get them from the docs.