Stand with Ukraine flag
Try it now Pricing
Trendz Analytics
Community Edition Professional Edition Cloud Edge PE Edge IoT Gateway License Server Trendz Analytics Mobile Application PE Mobile Application MQTT Broker
Documentation > Prediction > Overview
Getting Started
Guides Installation
On this page

Build prediction model for time series data

Trendz has built-in instruments for time-series prediction in few clicks. All required work, like data filtering, normalization and model training performed in the background. You can enable prediction for any fields including calculated fields. Time series prediction unlock a lot of insights that can be extract for your data. Here is a small list of questions that can be answered with Trendz time series prediction:

  • Forecast energy consumption for the next quarter or year
  • Predict when the next maintenance should be scheduled
  • Predict when the next failure will occur
  • Forecast manufacturing KPIs and how they are affected by current system state
  • Forecast remaining time till fuel tank will be empty

 

Prediction Models

Trendz implements different multivariable and univariable ML models for timeseries prediction:

  • Fourier Transformation - dissects a time series into its frequency components. Its prowess lies in discerning cyclic trends and seasonal patterns entrenched within the data, contributing to the efficacy of forecasting exercises.
  • Prophet - a forecasting paradigm devised by Facebook, meticulously crafted to handle time series datasets accentuated by pronounced seasonal patterns and holiday impacts. This technique utilizes an additive framework to encapsulate trends, seasonality, and holiday-induced influences.
  • Multi Prophet - evolves from the Prophet forecast model, enabling the simultaneous prediction of interconnected time series. This becomes particularly advantageous when dealing with numerous interconnected variables that warrant predictive insights.
  • ARIMA - combines autoregressive and moving average aspects to anticipate upcoming values founded on historical observations. This approach adeptly accommodates trends and seasonal variations ingrained within the dataset.
  • SARIMAX (Seasonal ARIMA with Exogenous Variables) - extends the ARIMA methodology by assimilating supplementary exogenous variables, capable of influencing the predictive outcomes. This model comes into play when external factors interplay with the time series, rendering it a valuable tool for enhanced forecasting accuracy.
  • Linerar Regresssion - a fundamental statistical method, comes into play to anticipate a dependent variable’s trajectory grounded in one or more independent variables. This technique establishes a linear connection between variables, forming the foundation for making forecasts.
  • Custom Model - you can write our own multivariable time series prediction model using any Python libraries. In this case you provide a model sources and Trendz is responsible for inserting input dataset from ThingsBoard and process forecasted output.

Simple Prediction

image

In this example we monitor vibration of pumps on water station. We want predict when vibration will rich critical levels so we can schedule preventive maintenance for the pump and avoid unplanned downtime.

  • Create line chart
  • Add Date(RAW) field on X-axis
  • Add Pump name field to Filter section and select required Pump
  • Add Pump vibration telemetry on Y-axis

Now Our Chart is ready for making forecast:

  • Enable Predction check box for forecast field
  • Select FOURIER_TRANSFORMATION as Prediction Method
  • Set Prediction Unit to days
  • Set Prediction Range to 2 - we make forecast for the next 2 days
Configure prediction for energy consumption on line chart
Configure prediction for energy consumption on line chart
Select prediction model and forecast horizon
Select prediction model and forecast horizon

Now we can press ‘Build report’ buten and check how our forecast looks like.

Forecast resource usage

In this example, we forecast how much energy would be consumed by teh building during next year. We have Buildings, Apartments and each Apartment has energy meter. We need to aggregate telemetry from sensors on building level and then build a forecast.

  • Create Bar chart
  • Add Date(RAW) field on X-axis
  • Add Building name field to Filter section and select required Building
  • Add Energy Meter energyConsumption telemetry on Y-axis
  • Enable Prediction check box for energyConsumption field
  • Select LINEAR_REGRESSION as Prediction Method
  • Set Prediction Unit to years
  • Set Prediction Range to 1 - we make forecast for the next year
Resource usage foreacast for the next year
Resource usage foreacast for the next year
Select prediction model and forecast horizon
Select prediction model and forecast horizon

Compare prediction with historical data

If you want to see how trained model looks like compared to historical data you need to enable Show Historical part checkbox in field configuration. Once enabled, predicted dataset will contain historical part predicted with the same model.

image

Next Steps