Statistics Enrichment Camp 2019

Final Analysis

Data Quality

There were many data entry errors. Some were perhaps my own fault, for not clarifying what to put as the Name. Others were surprising, for instance, entries of 100kg and heights of 20000m. We managed to obtain measurements from 62 individuals.

Baseline Measurements

## # A tibble: 2 x 2
## # Groups:   gender [2]
##   gender     n
##   <chr>  <int>
## 1 Female    18
## 2 Male      41
Try to obtain mean HR from interpolated object.

Compare Baseline Through Plots

When we inspect the points where HR is close to 150 and below 50, we clearly see something wrong with the heart rate sensor or the capture. Hence we will remove those points from the next plot onwards. We try one more plot, of BMI, and we add in an indicator for gender. There doesn’t seem to be much association with gender/bmi.

Comparing Sports and Doing Math

Spread the dataset and create a boxplot. If we had had time to obtain the full data properly, we would have had more power to make a better comparison (adjusting for individual), but we have too many individuals with only one or the other. The plot agrees with what we saw in the LT; the elevation (relative to baseline) is greater when working on the math problem, when compared to the elevation from watching sports.

Time to Recover

First, we filter out those we cannot use. The two metrics did not correlate highly enough. There has to be another round of data cleaning before they can be used properly. Even after the above, we see a lot of changepoints detected at 5 seconds, due to poor initialisation of the devices.