6 – Applications of Data Analytics

In this blog I am attempting to capture some applications of Data Analytics across different businesses and domains. If you are totally new to this field of study, at the end of the blog you will get an idea of how important Analytics is in the current scheme of things and in next years.

Marketing:

Point of Sales data from the checkout counters of shops can be analyzed to work out customer behavior, product associations and selling trends. Promotions, special pricing & use of in-store displays can be planned accordingly. Analytics can also be helpful in studying the post promotional data to see what impact it had on sales and profits. Future marketing strategies can be refined in this manner.

BFSI:

Analytics can be used in Finance to direct investment decisions. For example, in stock market the price/earnings ratio, dividend yields and the trend over a period of time can be compared with market averages. This can be used to make buy / sell decisions.
In Banking, risk analytics can be deployed to predict the probability of a customer repaying loans or to classify customer as a potential payer / non-payer of EMIs. This information can be used to drive decisions around who should be given loans and how much.

Manufacturing:

Data from the manufacturing floor can be collected from sensors on different machines. These can be analyzed for optimum use of the machines. Alerts can be set up to call for maintenance or repair, that can ensure bigger repairs and more expensive replacements are avoided. Also, production continuity is maintained without stalling of any section due to machine repairs. Similarly, alerts and notifications can be used to plan in advance for machine replacements thereby avoiding outage times.

Supply Chain Management:

Demand Forecasting helps in managing the uncertainty in the business environment we live in. If the demands are precisely known, a forecasting exercise will never be needed. But that is hardly the case and hence the need for forecasting methods. Forecasting methods work with history data and predict what will be future demand.
Based on the information available around Demand, Order Cost and Holding Cost, the Economic Order Quantity can be calculated. EOQ can be used to place orders of the optimal size. The Inventory size maintained will never be too high to shoot up the Holding Cost, nor too low to shoot up the reordering cost or cause schedule delays.
ReOrder Points can also be optimally estimated using the information available about Demand, Lead Times required by the suppliers for different components and the variability involved in these two factors. This ensures the different parties in supply chain are optimally engaged in the production cycle, thereby resulting in good supplier relationships and best possible Turn Around Times.
Buffer Size or Safety Stock can also be optimally estimated based on the forecasts, to enable adherence to targeted Service Levels.

Healthcare:

Health care analytics refers to collection of data from four areas within healthcare; claims and cost data, pharmaceutical and research and development (R&D) data, clinical data (collected from electronic medical records (EHRs)), and patient behavior and sentiment data. Data Mining on this data can reveal trends and patterns that can drive future decisions.
Recently there has been news about how a cognitive computing analytics product diagnosed a rare form of leukemia in a patient by analyzing and matching the patient’s genetic data with thousands of papers and research institutes across the world. This is accomplished at a much greater speed compared to human doctors and possibly at a greater accuracy. This could well be the difference between life and death for a patient.

Retail:

Future demand for various products can be forecasted to stock them accordingly. This approach to optimize inventory can save costs by avoiding unwanted stocks and also improve customer satisfaction by avoiding out-of-stock scenario.
Demand and competition can be bench marked to set optimal prices for products. This way businesses can avoid overpricing that could chase away customers, and at the same time set the highest prices possible to generate most profits.
In online retail scenario, analytics can be deployed to identify what product a customer has searched for. Similar options that he is highly likely to be interested in, can be displayed. Also, the related products and accessories that go along with the products purchased or added to cart can be presented.
Even facial recognition and image processing are used by big retail stores to assess customer likes and dislikes, analyze their expressions towards different products and plan accordingly for the future.

Accounting:

Auditing firms use sampling techniques to verify a portion of the expenses, accounts receivables etc., draw statistical inferences about the whole balance sheet and conclude whether it is acceptable.

Information Systems:

Statistical information like number of users on the system, system down time, bandwidth usage at different times of the day are collected. This is used by system administrators to assess and manage the performance of networks.

Entrepreneurship:

Customers can be segmented by geographic, demographic, psychographic and behavioral characteristics. Perception maps can be created to divide products of a certain category across two dimensions. For example, automobiles can be plotted across luxury and value-for-money as two dimensions. Similarly food items can be plotted against taste quotient and health quotient. In such plots, businesses can visually attempt to see where the existing products in the market are clustered and where there are gaps. Accordingly choice can be made on which customer segment to target and where to position the new product in the minds of consumers. Marketing strategies can be designed accordingly.
Also, geographic and demographic aspects can be analyzed to determine which is the best location to open a new store.

Politics:

We have heard from recent elections across the globe that political parties heavily use data scientists to design their campaigns. Parties try to assess people emotions, likes & dislikes by demography and accordingly choose words for speeches that will sensitize the mass towards voting for them. They also choose what items in the propaganda will give best chances of victory.

Quality Control:

With today’s emphasis on quality control, a variety of statistical quality control charts can be used to monitor Production output and compare them against set standards. This will let the business know when corrections are required in production processes.

Sports:

The difference between a cricket match telecast 2 decades ago and a match happening today is quite obvious. We can see the head coach and various other trainers sitting with their computers and closely watching every second of action. They break the game down into so many small portions and look at every minute detail. How a batsman scores most of his runs, which shots does he frequently get out to, what kind of bowler does he usually get out to, where exactly to bowl to him and so many other factors are worked out. Similarly, the optimal action for a bowler, whom to use him against and in what intervals, what field placements to use for certain bowlers and batsmen etc are all systematically planned.
Hope that gets you excited enough for a more technical study of Analytics.
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s