MegaFon Tajikistan introduces EW DataFlow to work with big data

2024: Implementing EW DataFlow to work with big data

On September 13, 2024, the company Eastwind announced that it had "" MegaFon Tajikistan implemented EW DataFlow to work with. With big data the Eastwind product, the operator optimized the department, analysts big data which helped it optimize the time-to-market of analytical ML models.

According to the company, "MegaFon Tajikistan" - telecom a company provider and digitalization in. Republic of Tajikistan Launched cellular networks of 3G standards ,/and. 4GLTE 5G

For more than six years, the operator has been using the CVM solution from Eastwind, Mobile Marketing Platform (MMP) to personalize marketing, increase the conversion of advertising campaigns and strengthen subscriber loyalty. This work was previously based on the analysis of client data, which was carried out in MMP using the built-in analytical module.

When MegaFon Tajikistan created independent lines of business analysis, CVM and Data Science, Eastwind proposed to expand the MMP solution with an additional module - a platform for transparent and convenient working with EW DataFlow big data. As a result of the implementation, the vendor integrated the product with other MMP components and transferred the processing of big operator data to it.

To create an MLOps system, operators usually use a set of open-source solutions that adjust to the infrastructure of their company. Such a development takes from six months and requires the presence of narrow specialists in the team for each tool. The DataFlow platform helped us build the right MLOps system in a month. With the solution from Eastwind, we got a clear tool integrated with all the systems we need. For September 2024, DataFlow allows data analysts to concentrate on their tasks and not be distracted by the devOps component. told Dmitry Kostenev, Head of Data Science at MegaFon Tajikistan

The platform from Eastwind allows data scientists to carry out all operations with big data and calculations in the Hadoop ecosystem in one interface:

• integration with any analytical resources
• Python и Jupiter Notebook
• real-time quality control of data and algorithms
• automatic translation of Python model into industrial code
• setting up data processing pipelines from an external source to the end user
• Data Lineage visualization

When implementing the platform, Eastwind connects its own data command to the customer's processes. The vendor helps the client set up data processing streams, connect all sources and implement AI/ML models to create a deep subscriber profile. Eastwind also provides consulting services for the rapid implementation of big data-based business cases and provides technical support in 24/7/365 mode.

At the time of the start of the DataFlow implementation, the operator launched more than a hundred regular data processing tasks through the analytical module, the results of which were used for marketing campaigns. As a result of the migration, Eastwind transferred all calculations of analytical models to another platform, without stopping a single advertising campaign. Analysts of MegaFon Tajikistan have the opportunity to implement and optimize ML models.

During the migration process, the customer expanded the plans for using the platform and asked to take these changes into account in the already launched project. In the last month of implementation, the client, on the recommendation of the vendor, increased the number of servers in the Hadoop cluster from four to six. Such power will be enough to support calculations and to develop all operator calculations.

As of September 2024, DataFlow is in active operation with MegaFon Tajikistan. told Vladimir Borkovsky, product manager of EW DataFlow

MegaFon Tajikistan influences the development of EW DataFlow as a whole. Eastwind regularly discusses the functionality of the installed platform and considers in detail all customer requests - some of them become part of the product roadmap. This is how the vendor optimizes its solution based on the real needs of customers.