Explore

Experiment in our foolproof environment.

The SaaS version of Datapred provides data scientists with a comprehensive framework
for rapidly building time series solutions that scale.

Contact us

Fast and secure

Datapred Explore is the best option for quickly building robust machine learning solutions for time series.

  • Use our library of built-in predictive models to prototype fast and assess the potential of your modeling ideas.
  • Our data pre-processing tools ensure that you stationarize your time series, extract the relevant features, implement robust backtests… while avoiding the dreaded “future leaks”.

Comprehensive framework

Datapred is the most open and comprehensive framework for time series modeling today, offering you complete control and endless possibilities at every modeling step:

  • Integrate and/or aggregate any predictive model, from moving averages to neural networks.
  • Choose your aggregation or stacking procedure from our built-in library, or implement your own.
  • Adapt your batch Keras, Scikit-learn, TensorFlow or PyTorch algorithms to the specificities of time series.
  • Define and optimize your own cost function - however complex.

Datapred automates the tedious or tricky aspects of time series modeling, letting you focus on exploration and performance.


Collaborative prototypes that scale

Datapred’s graph structure greatly accelerates the transition from prototype to production:

  • Backend optimization (e.g. parallelization, checkpoints) is automated.
  • Bugs or bottlenecks are easier to spot.
  • Collaboration is a breeze - just agree on data shape and let team members work on their own modules. Datapred automates output savings for every module of the graph.

Support from world-class experts

Enroll our expert data scientists to up your time series modeling game and make the most of Datapred.

  • Our kickoff webinar will walk you through a realistic sales prediction and resource allocation scenario, and demonstrate Datapred’s main capabilities along the way.
  • With the Starter Pack, one of our data scientists will spend 3 days showing you the Datapred ropes, typically at the beginning of your subscription.
  • The Project Pack is for Explore users that want to dig deep into a modeling project - usually because of its promising ROI. One of our experts will spend 10 days with you over 3 months to help you reach maximum performance.

Features

  • Data sequentialization
  • Stationarization and filtering
  • Built-in models
  • Compatibility with ML libraries
  • Aggregation and stacking
  • Custom cost functions
  • Parallelization and distribution
  • Connectors
  • Graph structure
  • Performance monitoring
  • Checkpoints and backups
  • Continual improvement