After first observation of the data

[marvinmin]

Here is something observed from the data:

1. The data we need is in the folder Protocol (18 activities).
2. The kinds of activities are different from subject to subject. We'd better choose subject 101 or subject 108 for our project. Since these two have 16 activities and longer time recorded.
3. Or we can choose three activities to work on based on the duration of the activities recorded (from PerformedActivitiesSummary.PDF).
4. For the feature heart rate, there are many missing values due to different sampling frequencies of the monitors. We should not simply remove these missing values. We'd better fill these missing values with the nearby value or based on the timestamps to decide how to replace these missing values.

Please feel free to add more observations.

[marvinmin]

Switched to a new dataset.