Friday and Monday were part of the Easter holiday, so the weekend rest was long. Thankfully, the school calendar observed the holidays. On my part, I had considered that there were no holidays for me given that I desired to catch up with my backlog of concepts not fully understood. I had also planned to spend some time reviewing some concepts in project management – I had a pending professional certification exam on it. At the end of the holidays, I did not do much. I did not gather my heart together to do work, I merely did the basic minimum. However, I was happy that I got all the much-needed rest.
Tuesday 11th April, work resumed, and with it, schoolwork. Data Analytics was our first class. We traded tackles with correlation and regression. I had met this duo several times – twice or more during my undergraduate studies and once during my Master of Science degree. Therefore, it was easy to follow the discussion. We discussed that correlation shows the relationship between two variables. Here are the major highlights.
- The correlation coefficient tells the story of the nature of the relationship between the variables. A negative coefficient indicates an inverse relationship while a positive one indicates a direct relationship; additionally, a coefficient less than 0.5 shows a weak relationship while one that is equal to or greater than 0.5 shows a strong relationship.
- Statistical software calculate the Pearson Product Moment Correlation when the two variables are quantitative – have discrete, continuous, and measurable values example length of pods and height. They however calculate the Spearman Rank Correlation for qualitative variables.
- Correlation does not mean causation. In other words, that the variable ‘number of boxes bought’ correlates with the variable ‘number of vouchers given’ does not mean that the number of vouchers given caused the number of boxes bought, and vice versa.
- A major limitation of correlation analysis is that it does not have forecast value. It only traces the relationship between variables but does not lend itself to forecasting the future values of either variable.
It is these last two points that highlight the usefulness of the second concept we learned – regression analysis. This analysis, in addition to showing the relationship between two variables, indicates causation and has forecast value. We learn that regression is the basis of machine learning and pattern recognition. Our discussion led on to the two parts of machine learning: supervised and unsupervised machine learning. The first is based on linear regression (works with continuous variables) and classification (uses ordinal variables), while the second includes clustering and association.
Wednesday 12th April, we took the Nature of Human Beings Course with Dr. Kemi Ogunyemi. Our discussion was sprung from the two Moodle Class Discussion Forums, one for why the course is relevant to our MBA study, and the other for how the interplay of the intellect and the will with the sensitive dimensions of man. For the latter, we were to use evidence from three texts – C. S. Lewis’s The Screwtape Letters, Viktor Frankl’s Man’s Search for Meaning, and Mitch Albom’s Tuesdays With Morrie. For want of energy, I will not get into the details of the conversation, but the crust of it was that the human intellect and Will can help us respond to very painful and displeasing situations in a more hopeful and renewing way. We all have the choice to evaluate and interpret our physical realities and make commitments to a constructive and healing path.
