Impact of Analytics in Daily Life

I am always intrigued by the examples of Technology touching our daily lives. Whoever knows a little bit about IT industry, knows most of the projects are initiated or implemented to support businesses. Of course businesses for the most part, in turn, influence?daily?lives but when I get to know?about the implementation of technology?directly connects technology with human lives- I become almost ecstatic.

When we talk about data and analytics- we, the people related to technology, often tend to think about the?tools and various?programming techniques?to achieve the desired result.

Like what type of Analytics we are talking about here- is it descriptive, predictive, prescriptive analytics- what programming language (Python or R), what is the storage, and so on.

But what some of us might be failing to notice the compelling use cases these Analytics are used for and how it is?enhancing human society.?

Take a case of Cancer cell detection- Machine learning algorithm can try to predict which cells might?have been?impacted?by?Cancer?among?thousands?of?images taken?through a?diagnostic tool. It makes easy for a medical professional to focus only on those set of cells?or images?to further rule out?or confirm?the possibility?of cancer?for a patient.

Data is the foundation?

If?we?want to take few steps back and understand how all these come together-?I?guess this is the right?point of discussion?to do so.

Years of study and analysis deserve little more context?I?believe.

A Journey always starts with data. Now every human being is generating millions of data points in their?daily life.?Every digital touchpoint- accessing social media, google search, website visit, buying items online is a data point for data scientists to consider.?However, data is like dis-joined lego pieces that need data engineering- years of historical data or millions of sample data need to be brought together?from various sources, put into some kind of model so that valuable insight can be?extracted?from the data. It is like solving a jigsaw puzzle. Scattered throughout, makes?little?sense?while spread all over the place?but if you try to put a framework, find out some common pattern/trend-

voila, you have a compelling story to tell about the data.

Different types of Analytics and their use:

Descriptive:

Once you have data?put in a place by Data engineers ready to?tell you a story - next try to run Analytics on the data.

We have Descriptive analytics- which is all about analyzing?and interpreting?historical data. An ideal example?could be?a business may want to know, what could the reason for their sales to be down last year??Why the warehouses are out of stock last quarter? What is the profit margin Y-o-Y?

One use case is typically used in?Industry?when a Business is about?to launch a new product. Data scientists are asked to run a study where they aggregate Sales, Operations data with Marketing campaigns, the?budget?of sales promotion,?and answer business - which promotion channel yielded maximum ROI (Return of investment).?Sometimes this is called TPO (Trade promotion optimization). When they should invest or how the promotion budget can be optimized.

Let’s assume as per the study it turns out?the investment in?Digital media has?translated?maximum sales as compared to print media. The business will align its strategy and invest more in digital media than print or billboards for promoting an upcoming product.

The business has observed as much as 70% of the budget spent on promotion does not break even which means that goes down the drain. So of course this is a big concern or focus area to make the decisions right while a brand new product is being launched.?

Predictive?

This is the most used and comprehensive analytics. Starting from Weather forecast to stock recommendation, from demand forecasting for a retailer to traffic route suggestion- predictive analytics has become an integral part of our lives.

Predictive analytics tells us what is most likely to happen based on previous data set.

It combines technology, mathematics, and statistics to identify the likelihood of future outcomes based on historical data. The technology is mostly Machine learning which deserves separate mention later.

While we are fiddling in an e-retailer website and it tries to tell us what we might be interested in buying or combine?another?product with already added items in the cart- all are based on Predictive analysis algorithms. Predictive analytics promotes cross-sales opportunities?by studying search history/buying patterns, help businesses attract/retain/grow most profitable customers, fraud detection by analyzing behavior or shopping pattern, send offers, etc.

Machine Learning:

So the mastermind behind predictive analytics is Machine learning. ML provides statistical tools to explore/analyze data whereas the most common use is a?Recommendation engine. It provides recommendations like the best time to provide a personalized offer to a customer, during monsoon what items are going to be sold more, and which items are less. Well, all businesses have some hypothesis about all of these but ML accepts,?rejects, or reinforces?that hypothesis.

In the era of Data analytics where?Retail?giant like Walmart generates 2.5 Petabyte of data from 1 Million customers every hour,?NOT taking account of these data will be an unpardonable offense and go by just a gut feeling.

Coming back to ML, the developed algorithms go through?data training-?just like how you teach a child what is good or bad with examples.?Hence it is divided into?supervised, unsupervised, reinforcement learning?but I don’t want to talk about them?just now else it may become a?dry read.?

Let’s stick to the use cases- Machine learning algorithm makes life easy for businesses. In a factory where millions of products are being made in few hours-?through image recognition it can detect damages?which saves huge business losses, return/replace/refund cycles, it can?optimize the route for logistics movement,?it can even sort apples based on size while packaging them. How cool is that??

My favorite one is-?Last mile tracking.?The supply chain’s efficiency heavily depends on the last-mile delivery cost which contributes to the major overall logistic mobility cost.?Last-mile tracking is - when you see a notification of “Out for Delivery” till it reaches your hand.?For me, it is really surprising to know that the last leg of the journey incurs the most cost.?So machine learning algorithms can look into past purchases, how customer put their addresses, how long in past it took to deliver, and optimize the shipping process to reduce the cost.?

When I was reading about the devastation of?Hurricane Ida?brought to North America- I was thinking if a Retail customer wants to build a gigantic warehouse that will be less impacted due to?future?natural calamities?and also quickly refill stock cross country once the situation stabilizes- it got to take help of predictive analytics to zero out a location.?It cannot be based on just a few key stakeholders' calls - rather more thorough analysis, strategy, and recommendation would be required which will be Prescriptive analytics.

Prescriptive?

Prescriptive analytics typically don’t employ in day-to-day business. It is to solve more profound issues, bigger and badder?which need an enterprise-wide solution. More like CX-level challenges.

So?Prescriptive analytics is a statistical method to find the ideal way forward or action necessary for a particular scenario, based on data. How it is then different from predictive? The difference is-?whereas Predictive most common use cases are recommendations that you may or may not take based on your decision, Prescriptive is more like taking actions. It focuses on actionable insight rather than data monitoring. The outcome of predictive analytics becomes an input to prescriptive analytics.

Prescriptive algorithms run?simulations?and do an impact/benefit analysis of each possible outcome and suggest you the best one for you to act on.

An example could be:?during a Clinical trial, it will tell you which is the best patient category based on age, ethnicity, demography, and best testing methods for most successful results.?I am sure before the Covid vaccine went live similar prescriptive analytics ran.

The business has to pivot on Prescriptive analytics to make critical decisions for any long-term scenario.

We are still uncovering new use cases to facilitate businesses as well as mankind at large where Analytics will work for greater cause and?well-being of?society.

Use cases are abundant- last day when I was reading the newspaper I came across the young prodigy who succumbed to pressure and took his own life after making a post on social media.

Can we put such an algorithm where Sentiment analysis will send alerts to authorities when people show indications of mental health issues or show symptoms of suicidal attempt and we can prevent such cases before it happens? I am sure social media platforms are already in process of bringing those or?maybe a?few are already in place?as we speak. It has its own thread of legislature or legalities but I am certain in the coming future we can make this earth a far better place to live in by using technology and analytics.

Want to share few more use cases you are aware of? Please use the comment box.