Missing data

[devhci]

Dear @brinnaebent,

I came across your paper 'Investigating sources of inaccuracy in wearable optical heart rate sensors'. It was mentioned that https://physionet.org/content/bigideaslab-step-hr-smartwatch/1.0/ data was used in this analysis.

Upon download and viewing the datasets https://physionet.org/content/bigideaslab-step-hr-smartwatch/1.0/. it appears that the majority of the sensors have <50% data compliance i.e., the columns have NaN values. Apple Watch has Max missing data. I wanted to inquire:

1. Is the data used in the paper ''Investigating sources of inaccuracy in wearable optical heart rate sensors' the same as https://physionet.org/content/bigideaslab-step-hr-smartwatch/1.0/ or is this a limited version of the full dataset used in the paper?

2. Difference/Inaccuracy in the HR from the ground truth ECG is one thing but missing almost 50% of that data is something that puzzles me. If so, is not the nearly 50% missing data a bigger problem than differences in HR from the round-truth ECG?

Regards
Devender

[brinnaebent]

Devender, Can you send an email with this information to my email ***@***.***? Thanks, Brinnae

…

On Jan 2, 2023, at 9:17 AM, Devender kumar ***@***.***> wrote:  Dear @brinnaebent, I came across your paper 'Investigating sources of inaccuracy in wearable optical heart rate sensors'. It was mentioned that https://physionet.org/content/bigideaslab-step-hr-smartwatch/1.0/ data was used in this analysis. Upon download and viewing the datasets https://physionet.org/content/bigideaslab-step-hr-smartwatch/1.0/. it appears that the majority of the sensors have <50% data compliance i.e., the columns have NaN values. Apple Watch has Max missing data. I wanted to inquire: Is the data used in the paper ''Investigating sources of inaccuracy in wearable optical heart rate sensors' the same as https://physionet.org/content/bigideaslab-step-hr-smartwatch/1.0/ or is this a limited version of the full dataset used in the paper? Difference/Inaccuracy in the HR from the ground truth ECG is one thing but missing almost 50% of that data is something that puzzles me. If so, is not the nearly 50% missing data a bigger problem than differences in HR from the round-truth ECG? Regards Devender — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.