Data analysis, visualization and Model demonstration tool.
Data analysis, visualization and Model demonstration interface using Streamlit. It’s a part of H-Beacon: Smart Irrigation System project. Enables easy and quick sensor data analysis, visualization and trained models inference for any newly acquired data without repeated coding. Application is deployed on Heroku.
H-Beacon is the deep sequential neural network model that estimates soil humidity from the strength of the LoRa-beacon IoT signal. We are funded by Horizon 2020 EU funding for Research & Innovation.
$ git clone https://github.com/TomislavZupanovic/H-Beacon-App.git
$ streamlit run app.py
Variables plot in respect to soil humidity
Plotting variables for chosen sensor for any time frame and moving average to see any occuring patterns in data.
Correlation matrix
Descriptive statistics and correlations matrix for any chosen time frame.
Scatter plot
Scatter plot between any two variables with dropdown menu for choosing.
Data transformations and analysis
Data can be easily transformed (Logarithm, Squared etc.) to check for distributions with histograms, applying operations to analize stationarity and standard deviations.
Model inference
Choosing trained models to estimate soil humidity on any time frame, showing metrics, residual and error plots for model performance.