A comprehensive analysis of weak signal propagation within the WSPR network, focusing on the impact of distance, frequency, and power on signal-to-noise ratios. The project includes data cleaning, statistical analysis, and linear regression modeling to predict signal reception quality and understand the factors influencing signal propagation.
A comprehensive analysis of weak signal propagation within the WSPR network, focusing on the impact of distance, frequency, and power on signal-to-noise ratios. The project includes data cleaning, statistical analysis, and machine learning modeling to predict signal reception quality and understand the factors influencing signal propagation.
- Data Cleaning: Handling missing values and imputing numerical data.
- Statistical Summary: Summary statistics and outlier detection.
- Unique Values Analysis: Counting unique values in specific columns.
- Distance Analysis: Calculating average distance for signals with specified power.
- Frequency Analysis: Identifying the receiving station with the most transmissions in a specified band.
- Data Partitioning: Splitting the data into training and testing sets.
- Linear Regression Modeling: Predicting signal-to-noise ratio based on distance, frequency, and power.
- Model Evaluation: Assessing model accuracy and making predictions.
The data set consists of the following columns:
id
: Unique identifier for each reception reporttime
: Date and time of signal received (YYYY-MM-DD HH:MM:SS format)band
: Fixed designator for frequency band (values: -1, 0, 1, 3, …, 1296)rx_sign
: Call sign of station receiving signalrx_lat
: Latitude of receiving stationrx_lon
: Longitude of receiving stationrx_loc
: Grid square of receiving stationtx_sign
: Call sign of transmitting stationtx_lat
: Latitude of transmitting stationtx_lon
: Longitude of transmitting stationtx_loc
: Grid square of transmitting stationdistance
: Distance between receiving and transmitting stations (km)azimuth
: Compass direction of signal received from transmitting stationrx_azimuth
: Compass direction of signal transmitted to receiving stationfrequency
: Receive frequency (Hz)power
: Transmission power (dBm)snr
: Signal-to-noise ratio of received signal (dB) in 2.5 kHz bandwidthdrift
: Reported frequency driftversion
: Receiver software versioncode
: Encoding type of signal
Performed data imputation on numerical columns to handle missing values.
Generated summary statistics for all numerical columns to understand data distribution and detect outliers.
Counted unique values in band
, rx_sign
, and tx_sign
columns.
Calculated average distance between transmitting and receiving stations for signals with power less than 30 dBm.
Identified the receiving station with the most transmissions on the 14 MHz band.
Randomly partitioned the data into training (80%) and testing (20%) sets using the createDataPartition()
function from the caret
package.
Generated a linear regression model to predict the signal-to-noise ratio based on distance, frequency, and power.
Evaluated the accuracy of the model using the testing data set and appropriate accuracy metrics. Discussed the performance and potential improvements.
Summarized the findings from the analysis and modeling. Provided insights into the factors affecting WSPR signal propagation and the effectiveness of the predictive model.
Contributions are welcome! Please submit a pull request or open an issue to discuss your ideas.