Refine your data into knowledge with Topological Data Analysis

Revolutionary approach to data analysis

Analysts are using queries to ask specific questions and receive specific answers. Received data are usually presented as tables.

+

Analysis of large data tables can be difficult. Modern approaches to data analysis help identify patterns and provide useful insights.

=

Get answers to questions you didn’t ask yet

Key features of TDA

Transform your data into a diagram

Visualise thousands of columns and millions of rows in one diagram. Understand complex and multi-dimensional data. Segment your customers and understand key characteristics of each group.

Get insights from the visualisation

See patterns and dependencies in your data. Be able to see how parameters depend on each other in a highly dimensional space. Forget bar charts and pie charts, analyse not 2, not 3 but thousands of parameters at once.

Answers for every industry

Challenge: Capitalise on internal and external long tail data

  • Identify anomalies in transactional history
  • Perform segmentation of transactional time-series and provided a key identifiers for each segment
  • Refine parameters for risk management systems and visualise current classification model
  • Analyse customer relationship data and provided detailed reports on customer segmentation

Advertising and marketing
Financial Services
Production and logistics
Telecom
Oil and gas
Entertainment

BigData Ready

Analyse gigabytes of data at the same time on the same diagram

Visualise hundreds of thousands of parameters and millions of rows

Supported SQL and NoSQL storages:

Better results

Result with Topological Data Analysis:

Analysis of NetFlix Prize Dataset. This visualisation represents 17,770 movies from the dataset. Topology calculation time: 5 minutes 32 seconds.

This data set is built on previous user ratings without any other information about users or films. The data of 100 millions ratings is presented as: <user, movie, date of grade, grade>

Data sets statistics:

  • 100,480,507 ratings
  • 480,189 users
  • 17,770 movies
  • 2.8 GB CSV file size

Download full movie map (3200x3200px)

Download full movie map with titles (17000x17000px, Size: 86MB)

Result with other dimension reduction algorithms:
Principal Component
Analysis
Isomap
Spectral Embedding
Locally-Linear
Embedding (LLE)
Hessian LLE
Local Tangent Space Alignment (LTSA)

Let's get in touch