This study presents the design, calibration, and in-field deployment of an IoT sensor network at a low cost for monitoring real-time urban air quality in Lagos, Nigeria, for diurnal and locational differences in Carbon-dioxide (CO₂), Nitrogen dioxide (NO₂), Methane (CH₄,) and weather parameters. The homemade system, constructed at ₦121,500, provides multi-gas and weather monitoring at 27–81% cheaper compared to mid- and high-end commercial sensors. Five hotspots with elevated risk—Oshosi, Iyana Ipaja, UNILAG dump site, Olusosun landfill, and Super Bus Stop—were assessed by morning (06:00–09:00) and evening (19:00–22:00) sessions. Nighttime levels of pollutants were always higher, with CO₂ highest at 790 ppm than the morning's 740 ppm, and CH₄ highest at 0.44 ppm ((Olusosun landfill). These increases were during times of elevated temperature (27.9-32.7 °C), humidity (68–79%) and lower atmospheric pressure (1003–1012 hPa), conditions that would have restricted vertical dispersion. High positive correlations (r ≥ 0.83, p < 0.05) existed between meteorological parameters (atmospheric pressure, relative humidity, and temperature) and the four detected pollutant concentrations (MQ135 index, NO₂, CO₂, and CH₄) during morning and nighttime sampling durations. The system maintained >95% uptime and gave data output within three seconds of data taking, and was therefore highly reliable and robust for tropical urban environments. This method offers an affordable, scalable model of continuous pollution monitoring and evidence-based urban environmental management for rapidly growing cities.