Digest

Data Strategy

Power your data-driven strategies.

Shapelets does more to empower your data science team.

Our platform is compatible with machine learning frameworks, try to integrate existing deep learning frameworks like TensorFlow or Keras.

Key Benefits

Optimized time series data

Engineered to optimize time series data analysis. Shapelets is packed with automated data preparation, data harmonization capabilities and native data type support for you to ensure your data is easy to read and free of redundancies.

Cloud scalability

We have optimized Cloud scalability in Shapelets. We are able to increase IT resources as needed to meet changing demands. Better storage, memory, and processing power ready for you.

Best in class algorithms ready to use

The solution provides preloaded algorithmics and is fed by state-of-the-art algorithms. Additionally, you can incorporate into the platform with your own algorithms or build on top of already available machine learning algorithms via building blocks.

Automated EDA

Digest provides EDA (Exploratory Data Analysis) guidance thanks to its AI. Thus, the work speed of the data scientist is increased, who in 3 lines of code can execute a model.

Optimized time series data

Engineered to optimize time series data analysis. Shapelets is packed with automated data preparation, data harmonization capabilities and native data type support for you to ensure your data is easy to read and free of redundancies.

Cloud scalability

We have optimized Cloud scalability in Shapelets. We are able to increase IT resources as needed to meet changing demands. Better storage, memory, and processing power ready for you.

Best in class algorithms ready to use

The solution provides preloaded algorithmics and is fed by state-of-the-art algorithms. Additionally, you can incorporate into the platform with your own algorithms or build on top of already available machine learning algorithms via building blocks.

Automated EDA

Digest provides EDA (Exploratory Data Analysis) guidance thanks to its AI. Thus, the work speed of the data scientist is increased, who in 3 lines of code can execute a model.

Run powerful data computations effortlessly

The Shapelets platform provides a provisioning system that helps manage infrastructure and calculations independently and reduce common resource availability issues in large organizations.​

Calculations can be scheduled, monitored, and managed from data apps.​

It also offers parallelized implementations of powerful, cutting-edge algorithms such as STOMP and the ability to parallelize external Python code for optimized execution on GPUs or compute clusters.​

This provides data teams with a powerful framework for expensive computations based on time-series data. It does not only ease of use and the ability to perform and monitor complex calculations. Our digest feature also includes powerful algorithms for detecting anomalies and motives, as well as the freedom to use external Python code and speed up its execution.

Main Components

Scheduled executions

You can now easily create timers and schedule periodic computations in the server. These Timer events can be used to trigger computations or update visual components or widgets in our intuitive UI. ​Thanks to this, data insights and visualizations can be automatically updated whenever it is needed. Consequently, if a task needs to run in large-scale production systems your data science team will need a stable, scalable, and monitorable solution.

Creation and monitoring of server-side functions

This feature will enable you to easily create deploy server-side computations and launch processes remotely. While in progress, the data professional will obtain partial computation results and feedback. It is a transparent and parallel execution that allows data scientists to cancel and relaunch computations very easily. In other words, these functions can be scheduled for periodic execution, run simultaneously with the data app, or executed upon events once registered. When running, these processes can be accurately monitored, inspected and terminated directly from the UI.

Shapelets’ Algorithmia

Shapelets platform incorporates an optimized implementation of state-of-art algorithms. Data teams will benefit from faster and distributed implementation of the Matrix Profile algorithm. It is a best-of-class algorithm for motif and anomaly detection in time series data analysis. It allows for efficient data analysis by quickly identifying motifs and anomalies in long time series data of both univariate and multivariate times series datasets. As the Matrix Profile is a vector that stores the z-normalized Euclidean distance between any subsequence within a time series and its nearest neighbour, it is a useful computation for motif and anomaly detection. Additionally, Shapelets digest features also incorporate the parallelized implementation of the STOMP algorithm that offers a computation complexity of O(n2) and allows it to overcome local minima.

Compatible with most Python libraries

Shapelets is fully managed using Python code and has no limitations on the Python libraries for developing specific applications within the platform. In other words, there are no version compatibility issues between Python libraries of interest and Shapelets’ framework. The benefits for this is that data teams will not need to learn how to use new tools and they can bring their own algorithms. In fact, you can turn standard algorithm implementations into optimal, distributed ones.

Scheduled executions

You can now easily create timers and schedule periodic computations in the server. These Timer events can be used to trigger computations or update visual components or widgets in our intuitive UI. ​Thanks to this, data insights and visualizations can be automatically updated whenever it is needed. Consequently, if a task needs to run in large-scale production systems your data science team will need a stable, scalable, and monitorable solution.

Creation and monitoring of server-side functions

This feature will enable you to easily create deploy server-side computations and launch processes remotely. While in progress, the data professional will obtain partial computation results and feedback. It is a transparent and parallel execution that allows data scientists to cancel and relaunch computations very easily. In other words, these functions can be scheduled for periodic execution, run simultaneously with the data app, or executed upon events once registered. When running, these processes can be accurately monitored, inspected and terminated directly from the UI.

Shapelets’ Algorithmia

Shapelets platform incorporates an optimized implementation of state-of-art algorithms. Data teams will benefit from faster and distributed implementation of the Matrix Profile algorithm. It is a best-of-class algorithm for motif and anomaly detection in time series data analysis. It allows for efficient data analysis by quickly identifying motifs and anomalies in long time series data of both univariate and multivariate times series datasets. As the Matrix Profile is a vector that stores the z-normalized Euclidean distance between any subsequence within a time series and its nearest neighbour, it is a useful computation for motif and anomaly detection. Additionally, Shapelets digest features also incorporate the parallelized implementation of the STOMP algorithm that offers a computation complexity of O(n2) and allows it to overcome local minima.

Compatible with most Python libraries