PNIPU Big Data Analysis Program

2025: Introducing the Big Data Analysis Program

The program of scientists of the Perm Polytechnic will make it easier for students and specialists to analyze large amounts of data. The university announced this on April 15, 2025.

In large companies, in order to service equipment, engineers and analysts often have to understand a huge amount of data. For example, to avoid breakdowns and repair equipment in time, you need to understand what affects the appearance of defects more strongly: temperature, vibration or wear? To do this, experts are engaged in data analysis using machine learning technologies, but these are complex areas that require deep knowledge of programming and mathematics. Scientists from Perm Polytechnic have created an e-book program that helps you easily understand the basics of data classification.

Certificate No. 2024690351 was issued for the program.

In equipment maintenance, engineering and analytics, specialists often have to work with large amounts of information. This allows you to predict when a particular equipment will fail and carry out preventive maintenance, reduce downtime and increase production efficiency. At production facilities, data is collected every second from thousands of sensors on pipelines, reactors, turbines and pumps. In large enterprises, their volume can reach from 1 to 100 TB per month (for comparison, this is approximately 1,000 hours of video or 310,000 photos). When working with such volumes of information, data analysts use special programs and approaches. However, they are very complex and require a high level of knowledge and skills from the user in the field of programming and mathematics.

To simplify big data operation and teach this, Perm Polytechnic scientists have developed a program that allows students and specialists without programming experience to master the basics of data classification and feature significance analysis.

The program is developed in the Visual Basic for Applications programming language and is an e-book. This allows it to run in any standard office packages, such as Excel Numbers, P7, Polaris Office,, etc. Open Office LibreOffice She uses a method that divides objects into two classes based on their features - this is called dichotomous classification.

Before full-fledged work, the program must be trained based on already known information. To do this, the user needs to independently divide the data that he enters into categories - "maximum" or "permissible" wear. The program will build a model that will take into account the relationships between characteristics and classes. Subsequently, the user will be able to enter new data, and the software will already automatically determine which class they belong to - whether this sensor will work correctly or not. Thanks to this, you can predict how the equipment will behave under certain conditions, predict possible accidents and prevent them, - said Yulia Bolshakova, training master of the dean of the Faculty of Applied Mathematics and Mechanics of PNIPU.

Such a program will be useful for maintenance and repair professionals, engineers and data analysts, as well as students and teachers of technical and economic specialties. Young specialists can use it to diagnose the condition of equipment (machines, sensors, etc.), predict failures and the need for repair work, students will help to better understand the basics of statistical analysis and classification of data, and for university teachers it will become an excellent visual tool for demonstrating how machine learning methods work. The program shows the principles of classification using real examples, you can change the parameters and immediately see the result.