
TelProgLab
A system for input data verification, TensorFlow forecasting model creation, and automation of training, forecasting, and reporting.
About TelProgLab
TelProgLab is software that enables extended input data verification and flexible creation of forecasting models using the TensorFlow library.
The application manages input and output data used in training and forecasting, supports neural network algorithm configuration, and defines forecasting, recurring, and reporting tasks.
The system architecture consists of the TelProgLab.server configuration layer, computational TelProgLab.runner instances, and a reporting module. Users access the GUI through a web browser with domain-based login.

Key features
An environment for the full forecasting process: from data source connection, analysis, and model training to result storage and reporting.
Links and data series
Define data providers, including TelWin SCADA and Folder, as well as daily or hourly data series.
Input data analysis
Check anomalies, empty periods, data series consistency, and behavior after projection to a different time resolution.
TensorFlow algorithms
Configure forecasting algorithms with input descriptions, output parameters, and detailed computation process settings.
Forecast tasks
Bind an algorithm with data series and a result repository, with actions for training and forecasting.
Reporting and computation
Report templates, communication with an email delivery bus, and support for multiple runner instances for demanding calculations.
Recurring task automation
A recurring operation defines an automatic, repeatable moment for starting a forecast task. The schedule can be used for regular model training, forecast generation, and report preparation.

Reporting and visualization
Forecast results can be presented as reports and charts, making trend analysis and comparison with real data easier.


Advanced data processing
TelProgLab provides tools for operations on input and output data: filtering, projection, conversion, and correlation analysis. This helps prepare consistent data for training and forecasting.
Initial operation implementation
Projection algorithms
Completion algorithms
Filtering algorithms
TensorFlow layer implementation
Support has been implemented for the following layers:
- SimpleRNN
- LSTM
- GRU
- Dense
- Dropout
- InputLayer
- Bidirectional
Final operation implementation
- Filtering
- Conversion
- Projection
Gallery
Explore the key work areas in TelProgLab (screenshots are available only in Polish).




TelProgLab service requests
Dear Customers!
To streamline the work of the Technical Support Department, please send service requests to:
Want to implement machine-learning-based forecasting?
Contact us to discuss how TelProgLab can be applied in your environment.
