In one place in SparkFun_Si7021_Breakout_Library.cpp, there is a while-loop with a delay after the Wire.requestFrom() call. That while-loop is not needed and that delay is not needed.

When the Wire.requestFrom() function returns, the I2C transaction has completely finished and the received data is in the receive buffer in the Wire library. The Wire.read() only reads data from that buffer.

Is there any chance to run in Attiny microcontroller?

At line 89 in SparkFun_Si7021_Breakout_Library.cpp the constant 175.25 is used for calculating the temperature but the datasheet of the Si7021 states 175.72.

As are the GND and 3.3V symbols.

at this line of the source code you use a literal 0xE0 which is what you defined TEMP_PREV to be. Suggest to use TEMP_PREV instead of 0xE0 for code readability

Hey Joel, sorry this review was delayed a couple days. Blame Halloween. Your layout looks good. Just a few little things.

1. Is the version number correct? [NO]
   comment: Should be v01 for proto - don't forget to fix in copper too.

2. Are any appropriate warnings (voltage conversion, sensitive I/O) labeled? [NO]
   comment: Are the SDA and SCL lines 5V tolerant? Either way, it'd be nice to add a note in the schematic

3. Does every part have a PROD_ID? [NO]
   comment: R1 R2 and C1 all need to be switched out for devices with prod ids

4. Are bottom side center ground pads 50 percent pasted? [NO]
   comment: Current paste is 100 percent coverage. Data sheed page 32: "A 2x1 array of 1.00 mm square openings on 1.30 mm pitch should be used for the
   center ground pad to achieve a target solder coverage of 50%."

I assume there's a sparkfun private repo that gets pushed to this repo. The library.properties indicates version 1.0.0, but 1.0.1 is in the wild- e.g., has been published. That makes it hard to track down possible problems and to know if pull requests would be honored here.

Is that worth it just so folks know they weren't mistakenly left out? I'm actually not sure if we have a standard for that, so maybe something to consider.

So we can add it to the Arduino library manager.

Regarding the code/library license, the main README, library README and library LICENSE file talks about "beerware" and "Use as you see fit", but the source code contains a GPL license, which is distinctly different.

Regarding the hardware, the main README, library README and main LICENSE talk about CC-SA 4.0, but the library LICENSE file talks about CC-SA 4.0.

I'm assuming the code license is GPL, as mentioned in the code itself, and the hardware is CC-SA 4.0?

at this line there is a delay(100) whilst as noted in the spec (and the comment in code) the max time for conversion of both RH and T° is 22.8ms.

waiting 100ms seems totally overkill and using delay() prevents other tasks from running and is not optimized for reading the temperature right after reading the RH since both are generated. suggest going with

`
// When not using clock stretching (*_NOHOLD commands) delay here

// is needed to wait for the measurement.

// According to datasheet the max. conversion time is ~22.8ms at full resolution

 if (command != TEMP_PREV) delay(23); 

`

Since there's not a ground plane, this isn't as huge of a deal, but we still don't want those potentially breaking the signal integrity.

Might be worth moving that to the right of the IC. This way if someone mounts it in a project with screw terminals, it's still visible (much nitpicking, I know).

Schematic symbols all say 3.3V, but you have a note for "VCC Rage:".
Board has 3.3V on bottom and VDD silk on top.

I'd double check and make sure you use the same label consistently. If VCC isn't applicable, delete that note.