Introduction
This is the 4th post in a series of 4 posts describing the types of data analysis(). This post is about Inferential analysis. Inferential analysis is based on inferential statistics which is a branch of statistics that primarily deals with going from data about the sample to insights about the entire population. An article that explains inferential statistics well is called Inferential Statistics: Definition, Uses by Stephanie Glen(the article can be found here).
Summary of Article
The article starts out by linking a video showing the difference between inferential statistics and descriptive statistics which is as the following: descriptive statistics tries to describe the dataset that you have with statistical measures(i.e. mean, median, mode, etc.), while inferential statistics uses the data from a sample and extrapolates it to the whole population. There are two major areas within inferential statistics which are: estimating parameters and testing hypotheses. Estimating parameters is when you take a descriptive statistic from the sample and then apply it to the whole population. Testing hypotheses, on the other hand, is where you actually test hypotheses through statistical tests to see, for example, if the improvement in productivity from using a certain app is actually due to that app or just random chance. If the improvement isn’t due to random chance then you can most likely apply it to the entire population. Then the article concludes with some information on a normal distribution and some statistical tests.
My Take
The article is a relatively short one but, when combined with the video that is embedded in it really gives a good high-level overview of inferential statistics. I feel it is similar to predictive analysis as in both types of analysis you are using data that you already have and trying to extrapolate it to something bigger(in inferential analysis it’s the whole population and in predictive analysis, it is the future). I especially found the section of the video about the areas of inferential statistics useful for data analysis. This is because using data from a sample and extrapolating it to the entire population can drive amazing insights for businesses and really help them be more informed and make more data-driven decisions. A really basic example is that if you owned a chain of malls and had data at a couple of malls on which types of stores people like in your malls then you could use both areas of inferential statistics(estimating parameters by if the number of people in the malls you surveyed didn’t’ like certain types of stores then you can extrapolate that to the whole population, and testing hypothesis by testing if a certain chain is liked in your mall) and then acting on those insights would generate a massive increase in the amount of revenue you earn.
Conclusion
All in all, this article was a good one. It really displayed the overarching concepts of inferential analysis well, while providing good, clear examples, while also sparking my interest in the field. I recommend you read this article when you get a chance(the article can be found here)!