A LoRa wireless communication architecture is proposed for censusing and data processing in Smart Cities, taking advantage of RLNC coding properties over packet segments to decrease redundant traffic in the system. At this point, the physical layer mechanisms ACM may not adapt modulation and error correction coding in time; residual errors cause packet losses at the upper layer that can be recovered with RLNC network coding. Thanks to RLNC coding algorithms in erasure channels, redundancy can be injected into packet transport, allowing receivers to recover quickly from packet loss, avoiding the costly RTT round-trip times incurred by relying on the traditional acknowledgment-based scheme. The system establishes a mapping of packet segments to a finite field (Galois field) scheme, which allows exploiting the algebraic properties of the system to exercise data recovery with information that has not been decoded