Is Data Science for me?

Anyone who has a task in hand to complete, a goal to achieve and has data available for the same, can use Data Science. We will approach this in an organized way. The idea is when a task in hand appears to be complex, vague, or overwhelming, break it down into smaller subtopics, arrange them in a chronological order and start analysing and looking at it in small packets. Once there are small or more precise things to address, challenges can be resolved effectively with better decisions. While we talk about decisions, Data Science has become a popular subject in today’s world which provides us with tools and techniques to understand the status quo, lookout for what can be improved based on the experience revealed by data. It helps us to take informed data driven decisions and thus guides us to further innovate and grow.

Consider your last 3–4 months grocery bills. Collect the information in a table format (This is your data). Look at the trend of the overall bill across the month i.e. the total sum of each bill. If there is an increasing trend, it is a good idea to further analyse this data and see if there is an opportunity to save in future (objective you want to achieve). Next is data exploration. Breakdown each bill across different types of grocery items that were bought e.g. rice, cereals, vegetables, fruits etc. Look at the trend of the bill amount for each of the grocery type and note the top 3 types contributing majorly to the bill. Further analyse the items within these types and identify the true contributor. You may find that a particular brand being bought in the near past is contributing to the increase in the bill. Decision could be to check for an acceptable alternative to control or reduce the bill amount. Hence, this is the way you can simply analyse data and take data driven decisions to achieve your objective. Below image is a data analysis illustration of the case we just discussed.

Data Analysis of Grocery Bills

When Data Science is applied on business data, we enter the world of “Business Analytics”. Quantum of business data is huge and more complex. Hence, we need sophisticated tools and techniques to analyse and use the information for business benefit. Let us take an example from a restaurant business which has been running for more than 15–20 years and is well established. They are interested in knowing how “Business Analytics” can be useful to them. Below is a step-by-step approach to address this curiosity.

About Data:

To apply science on data, we need data. Data from each sub-line of the business. For the above example, possible sub-line of business could be:

• Vendor data from whom the raw materials are sourced. This would also include quantity, frequency, pricing of the raw materials etc.

• Restaurant employee details. This would include their demographic details, skill details, designations, their work hours, monthly wages, bonus details etc.

• Restaurant operational details. Working hours, number of employees deployed, menu items available at different hours, table capacity, number of tables, number of customers being served at a given time etc. Additionally, electricity bills, water bills etc. will also help in understanding the operational spend.

• Customer details. Customer demographic details, customer feedback, customer bills, customer reviews from social media or internet etc.

• Quarterly or yearly financial statements of the business.

Above is a snapshot of data that the business may have. Even if business has not captured some data, we can always start with what is available and make provisions for capturing new data based on business priority.

As we see there is lot of data and if it is for 15–20 years, quantum is also huge. It is also important that the data collected is genuine and of good quality as this is what will be used further for analysis and decision making. Effective management of data i.e. maintaining its integrity, quality, confidentiality, security is owned and governed by Business Intelligence team (technical team), wherein they extract data from all sources, store them in data marts/ data warehouse/ data lake in a timely manner. This consolidated data which is information rich is further subjected to Data Science techniques to seek insights. Today there are also intelligent platforms available to manage the data journey right from sourcing to consumption.

Business Objectives:

Pick choose the business focus area and prioritize it. Broadly, any business focuses on 3 main areas:

1. Customer Satisfaction/ Customer Delight: Ways in which business analytics can help the restaurant business will be:

   i. Improve the Customer Satisfaction Score — Analyse and act upon customer review data.

   ii. Provide an offer to enhance Customer Experience — Analyse and act upon customer bill details.

   a. Personalized recommendation like a combo meal.

   b. Provide benefits to high value customers by gifting them a platinum /gold /silver card based on their past purchases.

2. Cost Optimization: Idea here is to reduce the cost/wastage. By analysing the vendor data and operational data, business processes can be streamlined.

   a. Need of staff can be regularized based on the capacity utilization in the restaurant.

   b. Vendor bills can be analysed for expensive raw materials and decisions can be taken to look out for acceptable alternatives that will be beneficial.

3. Expansion/Increase profit: There can be many options or strategies to achieve this.

   a. Include some discounts or provide lucrative deals during low business hours to attract more customers or increase sales.

   b. Make some arrangements for quick parcel deliveries to serve more clients during peak hours.

Workshops and brainstorming session between data science consultants and business leads, helps in identifying and prioritizing the business objectives.

Business Analytics:

Now that we have listed the business objectives, we can prioritize the same and start with Data Science activities (Note: The science that is applied on data is Mathematics and Statistics). First step is to perform exploratory data analysis wherein the data is summarised, relationships between data points is understood using descriptive statistics and visualization techniques (charts and graphs). It helps in understanding the current quantum of data, data quality challenges and the status quo of the business. By defining business relevant KPI’s (Key Performance Indicators), a dashboard can be prepared on the available data for management reviews and actions. At this stage itself, many of the business challenges gets identified on which actions and decisions can be taken. 

Further, based on the chosen business objective and insights from exploratory data analysis, machine learning or deep learning techniques can be applied on the data to get desired outcome e.g. building recommendations, customer segmentation, Natural Language Processing to analyse text data etc. These techniques are pre-defined algorithms packed in programming languages like R, Python etc. which are used to build models based on business specific data which helps business take data driven decisions. The job is performed by team of data experts and data scientists who ensure the built models are as per the desired accuracy and precision. Today, there are many trained models available for use which can be personalized based on the business objective.

So yes, data science can be used by a student to analyse his/her scores across different subjects over years, can be used by a home maker to manage household bills, can be used by working professionals to analyse their investments across different sectors and by any business to increase profit, provide operational excellence and ensure happy customers.