Besides the power constraints, you need also consider your range expectations. For instance, there are a few humidity sensors out there that do well in the mid-range of humidity senseing (from 15% to 85%); it is when you EXCEED this range that a 'better' (more expensive) sensor is required. I'm not up on my CO2 sensor, but if you are looking for variations from 'normal air' content, again, there are many in the 20% range. Lastly, you need to consider the distance for your remote sensor; the standard practice for remote sensing is the 4-20ma transmitter method, and, yes, there are 'off-the-shelf' solutions (4-20;CO2,Humidity,Temp. etc.) such as from' Vaisala' for (very approx.) $600 US/ channel.
If you want a carbon dioxide sensor module, try the GE Sensing T6613, it measures 0 ~ 2,000 ppm, the levels you would normally expect in a reasonably ventilated space. It has I2C, UART, and voltage outputs. But it has a trick up its sleeve, when using the digital output it will actually go to 20,000 ppm, albeit less accurately as it is outside the calibrated range.
However as the previous respondant mentions, the standard for longer distance measurement is either a 0 ~ 10 V signal, or a 4 ~ 20mA signal, the GE sensing T8100 range feature both these outputs, and look reasonably tidy too!
Try GE Sensing T8100 range, has voltage outputs (0~10 or 0~5) for Temp, Humidity, and Carbon Dioxide (CO2)