Analytics is like cooking
I like cooking. Mostly Asian food but occasionally some Western variations. I also like my job, which is largely evolving around data and analytics.
That’s why recently when I was asked to give a public lecture on data analytics, I marry the two; and here are how they look alike:
Data are like cooking ingredients. Without data there is no analysis; and bad data produce bad analysis. Different types of data need to treated differently, therefore we must know the characteristics of our data well. Data need to besourced, cleaned, stored, prepared, and used in proper manners, just like how you would handle your cooking ingredients. Data are codified differently for different purposes; and understanding those purposes would allow us to understand why certain data should be coded that way. Data manipulation and analysis techniques also vary according to the types of data, in the same way, you would cook your egg differently from the vegetables. For instance, survey responses on a discreet scale should never be added up or averaged as if the numbers are on a continuous scale. There is more to this analogy then just the “how to”. Many people are convinced that they cannot analyze their data because they don’t have the right data. Well, unless you really have nothing… analytics is possible even for small organizations and even if you only have a few data points. Remember this: big data is not a necessary condition for successful analytics. In many successful cases, organizations that were able to reaped the benefits of analytics often use only some of their internal data, marinated with some readily available external data like weather, economics, financial, seasons or time of the year etc. Some of my most favorite food use just a few ingredients!
Analysis is a combination of data, in a scientific way, to uncover the relationships among the different data elements. The same way as combining different ingredients when you cook; analysis is about uncovering (or testing) the relationships among different data. In some cases, you would need to analyze the data in a sequential manner (e.g. step-wise regression); in other cases, you may need to first build a comprehensive framework to model the inter-connected relationships (e.g. structural equation modelling). The key outcomes from an analysis is the establishment of the relationships among data elements, to the extent that certain conclusions can then be drawn. The good thing about performing an analysis these days is that analytical tools are becoming more sophisticated and powerful but at the same time cheaper and more user friendly. That got me thinking about my new air fryer: better, faster, healthier…
If data are ingredients and analysis is the combining and cooking of those ingredients, analytics is like serving a meal. Here I am talking about a meal from appetizer to entrée to main course to desert to the different drinks combinations to the table setting to the ambience, and most importantly, the people you serve. In analytics terms, data analytics is the end-to-end process of a series of analyses aiming at supporting decision making. When you are cooking up a meal, knowing who is coming and their likes and dislikes is often as important as the quality of the dishes itself. For analytics, switch thelikes and dislikes with the decisions that your audience need to make; if analytics can help them make better decisions, I am sure they will keep coming back for more. Secondly, while you may be serving a number of different dishes over the meal, those dishes need to go well with one another. They need to be “connected” in certain palatable ways. Likewise, analytics look beyond analysis in one area and sought to connect the dots across different departments or functions. The power of analytics is usually amplified when the right “dots” are connected. It is worth mentioning that, while there is nothing wrong to serve up a selection of unrelated dishes if a buffet spread is what you intended, just make sure your guests know what to expect.
So if you are tasked to lead an analytics project or function, just remember to treat data the same way you would treat your ingredients. Get the right tools for the analyses you need; and always start by knowing who your audiences are and what they need.
