I would like to share the results of my research with the community.
Polar IQ Compression — Research & Analysis
RTL-SDR devices capture radio signals as IQ data (In-phase / Quadrature), where each sample is a pair of bytes (I, J) representing the real and imaginary parts of a complex signal. These pairs have a natural geometric interpretation: they are points in a 2D plane.
Standard compressors (gzip, bz2, lzma) treat this data as a flat byte stream, unaware of its structure. The hypothesis behind this project was:
If IQ samples are naturally distributed along circles in the (I, Q) plane — as they are in many modulation schemes — then regrouping them by radius before encoding could expose redundancy that standard algorithms miss.
Status: 
Non-competitive vs standard algorithms
Conclusion: Polar coordinate transformation cannot overcome information-theoretic limits for general-purpose compression of IQ data.
Value: Rigorous documentation of a negative result, with complete mathematical analysis.
Link to GitHub repo: