Why should you use TDSS?

Scalable Infrastructure

Go beyond sampled analytics, run predictive modeling on complete data through our native Apache Spark Software distribution.

Prepare Data in Minutes

Use our automated data transformation layer allowing users to process variety of data at speed.

Production-Ready Deployment

Production ready server infrastructure reduces at least 50% deployment time through our dedicated cloud and private environments.

Easy Model Creation

Build stable models through our pre-built, auto-optimised grid search algorithms. No need to code.

Think Beyond Statistical Tools

Predictive Modeling Features TDSS
Run predictive modeling on production data
One-step model creation with no code development
Massively parallelized predictive models
Assemble and transform data in significant less time
Choose your deployment environment - Hardware Cluster,
Storage and Load Balancer etc

Much more than Business Intelligence


  • Automated and scalable predictive modeling.
  • Predict based on ML techniques

BI Tools

  • Data Analytics and Visualization Platform
  • Make inferences from existing data sets

TDSS - Flow

TDSS - Technology

Meet our killer tech. Production-ready distributed predictive modeling platform.

  • Automated accurate model creation through in-built hyper-parameter search and parallel model scoring.
  • Improve model performances with ease, based on feature engineering and ensemble techniques though pre-packaged transformational functions.
  • Save data transformation steps to schedule periodic runs on current data sets to create updated model data sets.

Feature Engineering

  • Data Transformation: Prepare data in minutes using multiple functions like binning, thresholder, window, aggregate, filter etc

Predictive Modeling Modules

  • Clustering: segment data into useful clusters for applications or as input to other models.
  • Regression: perform regression analysis using boosted tree models.
  • Classification: perform classification using boosted tree models.
  • Decision Tree: create decision tree through auto-optimised models.
  • Text Analysis: analyse and quantify large volumes of unstructured text documents using NLP mechanism.

Measurement Metrics

  • Operational metrics: capture cluster params, no of CPU cores used, memory params.
  • Performance metrics: RMSE, RSE, precision, recall, Fscore etc


  • Rest API Endpoints: request REST API endpoint to integrate client applications
  • TDSS Dashboard: visualise decision work flow and prediction results.

TDSS - Model Variations

Know our Varied Models. Scalable and Business Friendly Modeling Techniques.

Prescriptive Modeling

  • Identify drivers relevant for your business.
  • Discover business rules to segment and dive deep into your use cases.
  • Create decision trees and rules to understand relationships among business drivers.

Predictive Modeling

  • Predict incidences based on historical data and current scenarios.
  • Classify or label data based on identified patterns, which are learned from trained data sets.
  • Forecast scenarios by identifying right set of variables that affect future outcomes.

TDSS - Deployment

Get Production-Ready Infrastructure. Flexibly deploy on Cloud or On Premise.

Deploy on Cloud

Choose your dedicated deployment environment - CPU clusters, Storage, Load Balancers

Deploy on Premise

Deploy our software distribution in your own Spark cluster environment.

Deploy anything

  • Assemble all your data in one single place to perform data cleaning and processing using our powerful transformational functions.
  • Create multiple models and make them available for real-time predictions.

Deploy anywhere

  • Create production ready environments on Amazon cloud.
  • Utilize your existing Hadoop or Spark Clusters as execution environments.

Production-Ready Predictive Services

  • Production ready hosted models.
  • Re-train models on updated data set periodically.
  • REST API based predictions to integrate into existing applications.

Access to each team member

  • Train new models and perform ad-hoc analysis.
  • Train models on larger dataset periodically.
  • Compare results across models and choose best decision path.
Request Demo