I’m really going to confuse Twitter with the title of this blog post. My apologies in advance to the entity that owns the Twitter handle, “@CORRELATION”. Seems like a dead handle, anyways.
So I haven’t posted a technical post in about a month. Such is life sometimes, but I’m going to post one today as I’ve removed it from my Kscope presentation. Voila – blog for you. It’s short, but it’s pretty sweet.
When I was starting my Ironman training database, I had intended upon seeing if there was a strong correlation coefficient between the calories I burned in a workout (guess you could say, time and intensity) and the time of day at which I worked out (ie: 6AM or 6PM). However, after reviewing my own data for 2 month’s time, I abandoned that grandiose plan as, well, I’m just plain a morning person. I rarely workout in the evenings, so *of course* the correlation coefficient would be strong.
But I digress…
Using my own data as an example, I took my workout from January 7th to see if there was any strong correlation coefficent to the calories burned (822) and the time I started working out (5:40AM). Note that some of the items are greyed out not because I don’t want the world to know the data, but, rather, because they will be presented at Kscope.
The SKIPNONE means all values will be included in the calculation.
The next section (with the RANGE’s) are what I am comparing – TimeOfDay1 and “CalsBurned”. Both for January.
When this calc is run, I get a coefficient of 0.99316539. The range only goes from -1 to 1…so, yes, the is a strong correlation. I could extrapolate this example to the past 6 months, but the correlation wouldn’t be much different. I’m just a morning person! …Don’t shoot me…