Big Data Analytics in Business

Lecture 3: Big Data Analytics in Business (1)

• Outlines business intelligence, types and examples of big data analytics, and technologies like recommendation systems and cookies.

Business Intelligence (BI)

• Business Intelligence tools have been used for decades.
• It consists of applications/tech for gathering, storing, analyzing and providing data access.
• BI provides historical, current and predictive business operation views.
• Business Intelligence is compiled using Big Data.

Revenue from Big Data and Business Analytics

• Big data and business analytics revenue, worldwide, from 2015-2022 in billion USD.

Big Data Analytics Across Industries

• Big Data Analytics are applied in healthcare, telecom, finance & automobile industries.
• Big Data Analytics are used in insurance, government, education and retail.
• There is a focus on big data in business.

Types of Big Data Analytics

• Descriptive Analytics answer, What happened? and analyzes the past.
• Predictive Analytics answer, What could happen? and forecasts the future.
• Prescriptive Analytics answer, How can we make it happen? and suggests a decision.

Descriptive Analytics Explained

• The most traditional type of business analytics.
• Descriptive Analytics makes up the majority of all current business analytics
• It looks at past performances.
• It uses data mining to look for reasons behind past successes or failures.
• Descriptive analytics employs data aggregation and data mining.
• e.g. average weekly sales amount?

Predictive Analytics Explained

• Statistical models and forecasting techniques are used to understand the future.
• e.g. Forecast smartphone sales demand next year.
• It provides companies with actionable insights based on data.
• Offers estimates on the likelihood of a future outcome.

Prescriptive Analytics Explained

• Goes beyond predicting outcomes, suggesting actions to benefit from predictions.
• This shows decision-makers the implications of decision options.
• Prescriptive Analytics anticipates what, when and why something will happen.
• It suggests decision options by mitigating a future risk and illustrating its implications.
• Can improve prediction accuracy and offer better decision options.
• e.g. Google self-driving car.

Google Self Driving Car

• Big data is used to operate Google self-driving cars.

Big Data in Aviation Industry

• Big data examples in aviation include flight times, social media activity and in-flight food preferences.
• Number of people traveling is also useful data.
• Usages for big data include aircraft maintenance and inventory, fleet, operations, pilot/crew/staff management
• Benefits of big data usages include reduced cost, smarter maintenance, safety, and improved service

Aviation Industry Cases: United Airlines

• To encourage loyalty, United Airlines analyzes over 150 variables in each customer profile.
• As an example, passengers over 50 who like to take domestic airlines would be analyzed.
• The analysis measures previous purchases to customer preferences to generate tailor-made offers.

Aviation Industry Cases: British Airways

• To get to know their customers British Airways uses an intelligent 'Know Me' feature to provide personalized search results
• Relevant offers are provided using in-depth data analysis

Aviation Industry Cases: EasyJet

• To deploy artificial intelligence EasyJet invested in artificial intelligent determining seat pricing.
• The AI determines seating prices automatically depending on demand.
• EasyJet analyzes historical data to predict demand patterns up to a year in advance.

Aviation Industry Cases: Southwest Airlines

• In-flight intelligence is used to team up with NASA to continually improve airline safety.
• An automated system has been created with NASA to generate data to flag anomalies and prevent accidents.

Aviation Industry cases: Cathay Pacific

• To stay competitive and enhance passenger experience Cathay Pacific uses big data/IoT in flight planning
• They also use it for fuel optimization and preventive maintenance.
• The aircrafts are equipped with flight monitors, capturing data on safety analysis and corporate training
• Working with Rolls Royce they optimize jet engine operations.
• Enhance customer experience through combining sales/marketing data to offer travel options.

Big Data in Retail Industry Cases: Alibaba

• Big Data is influential in redefining many industries in the cloud computing era.
• $84.5B in online transactions during Singles' Day sale in 2021.
• Alibaba leverages Big Data analytics to enable cross-selling and up-selling at TMall and Taobao.
• Shopping habits, payment and credit history, demographics, search preferences, social networks and personal interests are analyzed.
• Real-time online data is used to predict consumer wants for recommendations.

Recommendation Systems: Definition

• Input is feedback of likes and dislikes.
• Output is recommended items based on the feedback.
• Examples include Netflix/Youtube movie recommendations, Amazon "customer who bought this" section, Spotify and Google.

Recommendation Systems: Approaches

• Collaborative filtering and content-based filtering

Content-based Filtering

• Content-based filtering builds recommenders based on item attributes.
• Item attributes for movies are genre, actor, director, producer, and hero.
• User taste identifies attribute values/weights, provided as input.
• Purely domain-specific, recommends movies in western genre to those who watch cowboy movies.

Collaborative Filtering

• Recommends items based on similarity measures.
• Items that are preferred by similar users will be recommended to another user.
• This technique is not domain-specific, unlike content-based filtering.
• Two types are user-based, and item-based.

User-based Collaborative Filtering

• Algorithm based on similarities between users.
• It's built on the idea of user sharing preferences.
• Recommends Movie2 to User1 from the movie taste ratings of User3

Item-based Collaborative Filtering

• Based on similarities between items.
• Runs on the idea of user having a preference for similar items.
• The P value is compared against threshold to determine a recommendation.

Item-based Collaborative Filtering: Recommender

• An item-based recommender needs preference data and a notion of similarity between items
• Item filtering systems is much better than user-based.

Recommendation System: Limitations--The Cold Start

• Collaborative filtering systems use available data from similar users.
• A content-based filter can be used first.

Recommendation System: Limitations--Scalability

• As the number of users grow, the algorithms suffer scalability issues.

Recommendation System: Limitations--Data

• Lacks accurate inputData because humans are not perfect at providing ratings
• User behavior is more important than ratings.
• Item-based recommendations provide a better answer in this case.

Cookie

• Small data files written to user's hard drive when browsing a website or filling out a form.
• Cookies record browsing activity, and arbitrary pieces of information that was previously entered.
• Used to re-access websites at a later date.

Cookie - Guidance

• Provide guidance about data collection and use
• Cookies can show previous/last page visited, and list similar items for customer to purchase

Big Data in Retail Industry: Chow Tai Fook Jewelry

• A huge bulk of consumer/product data is collected from different channels e.g. retail, e-commerce.
• Individual customers tastes, buying habits are identified.
• This is used to predict business performance.

Big Data in Retail Industry: Li & Fung in China

• Set up experiment shopping mall in Shanghai (Explorium.)
• Using technology to understand better through preferences
• different state-of-the-art technologies and big data analytics track/identify shopper behavior.

Banking & Finance: Hang Seng Bank

• Accepts online personal loan applications without income/address proof.
• Analyzes customer behaviors and creditworthiness
• Decides whether to approve the loan application using big data.

Banking & Finance: Citibank

• Over 200 million customer accounts globally.
• Analyzes data and targets promotional spending.
• Scans transactional records to spot anomalies such as incorrect, unusual, or fraudulent charges.
• Implemented successfully in customer retention and acquisition

Banking & Finance: WeLend

• WeLand approves loans in 1.7 seconds
• Determines creditworthiness with proprietary big data.
• Collect wider variety of data e.g. phone model, apps and network.
• Loan delinquency rate: 1% in HK.

Insurance: HSBC Insurance

• They understand customer analytics digitally
• Smarter, cheaper products are made.
• More products and manage risks, operate at lower cost and understand risk.
• Predict illness and accidents and price their products accordingly

Insurance: MLC Life Insurance

• Wearable tech encourages healthy lifestyles.
• 17% sign up for better health and reduce life insurance premiums.
• It provides product innovation and improved customer experience
• Also gives better engagement with customers.

Insurance: ZhongAn Online Casualty and Property Insurance

• ZhongAn is China’s first online insurer.
• Big data platform for companies along the car value chain.
• One-stop shop for buyers/owners.
• Offers better rates based on mileage, and is driving behavior for better pricing.

Hospitality Industry: Marriott

• It uses unstructured/semi-structured data - weather reports, and local schedules
• This enables them to forecast demand and value for each individual Hotel room all year.
• These are for optimum efficiency.

Hospitality Industry: Starwood Hotels and Resorts

• 1200 Hotels around the world
• By examining local analysis of data room pricing is optimized.
• An increase of revenue by 5% is caused.

Big Data Applications in Business Sectors

• Airlines: malfunction, safety, maintenance, cost and customer service
• Retail: optimize business
• Banking/Finance: differentiate fraud, and block irregular transaction
• Insurance: risk assessment, market, customer insights and experience
• Hospitality: price, data decisions and campaigns

Big Data Around You

• Communication
• Retail
• Industrial
• Vehicle
• Medical