CODEWITHSUNDEEP
androidwear-os
Ha Le
Ha Le

Reputation: 31

How to interprete the data from Pixel watch 2 EDA sensor?

I recently tried the Pixel Watch 2 because of its various health-related sensors, including EDA (Electrodermal Activity) sensor. I was able to register and receive events from the sensor, but I'm not sure how to interprete the results. Do anyone know if Google released any documentation about the Pixel watch 2 EDA sensor?

Examples of the events I received from onSensorChange() - the first entry of the row is the timestamp of each event:

1697814182743,[370173.0, 501147.0, 624072.0, 741215.0, 552115.0, 368724.0, 206038.0, 54419.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814182783,[213591.0, 371112.0, 510976.0, 640507.0, 460985.0, 283599.0, 124945.0, -22683.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814182824,[140744.0, 302128.0, 444450.0, 576944.0, 398948.0, 223772.0, 67359.0, -77733.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814182863,[86277.0, 249451.0, 394809.0, 529139.0, 352628.0, 179117.0, 24103.0, -119838.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814182903,[47015.0, 211219.0, 357679.0, 493353.0, 317703.0, 145919.0, -8466.0, -152088.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814182943,[15727.0, 180879.0, 327807.0, 463539.0, 289184.0, 116765.0, -36701.0, -179673.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814182983,[-11532.0, 153490.0, 301166.0, 438435.0, 264864.0, 93136.0, -59733.0, -201858.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814183022,[-34407.0, 132193.0, 280564.0, 418005.0, 245076.0, 73938.0, -77536.0, -219441.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814183062,[-50556.0, 116237.0, 264595.0, 402568.0, 230923.0, 60425.0, -91601.0, -232762.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

1697814183102,[-63904.0, 102562.0, 250841.0, 388963.0, 217071.0, 46788.0, -104393.0, -245124.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

I have tried searching through Android documentation and posts about pixel watch 2

Upvotes: 3

Views: 731

Answers (1)

javiercmh
javiercmh

Reputation: 9

I have been working with Empatica which reports data in a similar way: a timestamp followed by a sequence of measures. Based on this, one would assume these are time series and you could assign a timestamp to each data point if you know the sample rate of the sensor.

However, there are some problems with that assumption in your case. If you say you have skin conductance measures, these values should not be negative. Second, the trailing zeroes in all your batches look strange. In the context of skin conductance measure, it should be uncommon to get readings of absolute 0 which would mean no conductance at all. Third, it is safe to assume that these timestamps are in milliseconds (so they would have been taken in October, 2023). Given that you are getting 16 samples every 0.04 seconds, this would translate to 400 samples per second, which is very unlikely knowing that slow sampling rates (<10 Hz) are enough for EDA (source).

I recommend you to check the API again, because you might have been working with the wrong data.

Good luck!

Upvotes: 0

Related Questions