After being introduced to the ins and outs of Business Analytics, this week we continued the discussion by reviewing key concepts pertinent to the topic. The following set of videos revolved around different statistical concepts that one will see often if they decide to be a Business Analyst in the future.
The first set of videos discussed thoroughly about descriptive statistics, specifically Measures of Distribution and Measures of Variation. These are commonly seen at the start of any Statistics subject. With these measures, you can describe your dataset based on how the values are distributed (mean, median, mode, and the like) and how values vary (range, standard deviation, variance, and the like). Moreover, we were introduced to contingency tables that are extremely helpful in looking at frequencies of chosen variables in a dataset and visualization techniques that make data easier to understand using visual aids. These measures and tools make business data easy to digest, especially for stakeholders who would only want to see the information they need and not the raw data.
Following those videos, we moved onto the different types of distribution and sampling basics. Normal distribution tells us most values in dataset are near the mean. Moreover, if we have normal distribution, we also have asymmetrical distribution. This type of distribution is skewed as values do not appear at the same frequency, which also goes for its mean, mode, and median. To assess your datasets better, the concept of kurtosis is relevant as you can easily see the outliers that exist. As to sampling basics, for an individual to have a dataset to analyze, knowledge about population and sampling is important. When the population of your target data is too large for you to use, you are free to use different sampling methods to lessen this number. Examples of these methods discussed are random, cluster and stratified sampling. As analysts, we should have a keen eye on what specific methods to use depending on what is needed.
In the next section, we were introduced to Bivariate Data and Correlation, Information Theory and Entropy, Analytical Reports, Automation, and Regression Analysis. In being a Business Analyst, you are most likely to meet situations that will allow you to compare or find the relationship between the two quantitative variables. Evaluation with two variables can be done with Regression Analysis, where you can find association between a dependent and independent variable or Correlation, where you can see the relationship and how strong the relationship is. In Correlation, the strength of the relationship can be seen visually through a scatter plot. Furthermore, Information Theory and Entropy play a key role in creating Analytical Reports. In crafting reports, it is important to know what information should be included. Using the concept of Entropy, uncertainty becomes a good thing as these details give us more information and should be included in a report. However, moving forward as analysts, we should use other variables to reduce this uncertainty.
Lastly, choosing which statistical tool fits a specific problem was discussed. It is important to remember these three questions in figuring out the proper tool. First, what type of data is given? How many samples are there? And lastly, what is the purpose of the test? Answering these three questions can give you a guide on the right tool to use. Additionally, more statistical concepts were presented. These include hypothesis testing, statistical errors, and binomial probability distributions.
These are all theories and concepts a Business Analyst should know or be familiar with. Given how important evidence-backed decisions are needed in businesses, and how competitive the market is, knowing how to work around or manipulate data to gain insights really is non-negotiable when you are an analyst.
Posted using Honouree