About ETNA

ETNA is an open-source tool that is currently available in the programming language Python, but will in the future become accessible to a wider audience of users.

It is an open-source tool that uses data uploaded by users and gathered from public sources to conduct analyses, verify hypotheses and generate forecasts. Among its capacities, ETNA can be used to determine those factors that most affect a company’s profit, forecast demand for specific products, and calculate next year’s budget.

ETNA uses machine learning algorithms and neural networks, while ensuring total data security by storing downloaded information directly on the user’s computer and not sharing it with third parties.

How Tinkoff uses ETNA

Tinkoff uses ETNA to plan its hiring process, determine its need for office supplies, and assess how various business processes affect one another.

To give one example, ETNA helps forecast demand for cash withdrawals and deposits at specific ATMs in order to determine their replenishment schedule and analyse whether or not they need additional cash.

Experts at Tinkoff are also using ETNA to determine whether it is possible to predict the future spending of Tinkoff customers across various categories of goods, in order to help them plan their budget more carefully.

How ETNA was created and how it differs from international equivalents

Similar tools have been developed by a wide range of companies across the globe, including research institutions, startups that sell these technologies to businesses, and IT giants (Facebook, Amazon, etc.). The latter developed tools similar to ETNA for in-house use, before scaling them into standalone B2B products.

Tinkoff’s Centre for AI Technologies developed ETNA for the same reason – to handle in-house tasks related to the Tinkoff ecosystem. After discovering that none of the tools available internationally were flexible enough to meet all of the company’s business requirements, company experts decided to make their own and adjust it to Tinkoff’s specific needs.

ETNA outperforms similar international solutions due to both its user-friendly interface and its wide range of data processing methods and forecasting models.

How ETNA became an open-source project

Tinkoff decided to bring ETNA to the market after seeing how effectively the tool was able to solve the internal problems of a leading fintech company. ETNA is now an open-source tool, which means that anyone can customize it by changing its parameters, adding or removing certain functions, or modifying its software to better suit a particular industry.

To get a closer look at ETNA and start working with it, please visit our website, where you can test the tool’s capabilities through a series of examples that open directly in browser.

Pavel Kalaidin, Director of Artificial Intelligence at Tinkoff:

“ETNA is the first open-source project from Tinkoff’s Centre for AI Technologies. It is an incredibly flexible tool for data processing that includes all those things we couldn’t find in existing international solutions – namely, a convenient user interface and a much wider range of analytical methods and forecasting models.

At present, this tool can be used by developers and analysts familiar with Python. But we know the importance of open-source software for the scientific and business community, and are ready to further our contribution. We plan to make ETNA accessible to a wider audience in the future”.

ETNA is a proprietary product of Tinkoff’s Centre for AI Technologies. The Centre boasts over 20 teams and more than 200 machine learning experts. Products developed by the Centre include Tinkoff’s voice assistant Oleg, its VoiceKIT technology for speech synthesis and recognition, and predictive models for the company’s lifestyle services, content-based projects, and investment products.