Big data is now being generated all around us. So what? It's the applications. It is the way in which big data can serve human needs that makes it valued. Let's look at a few examples of the applications big data is allowing us to imagine and build. Big data allows us to build better models, which produce higher precision results. We are witnessing hugely innovative approaches in how companies market themselves and sell products. How human resources are managed. How disasters are responded to. And many other applications that evidenced based data is being used to influence decisions. What exactly does that mean? Here is one example. Many of you might have experienced it, I do. Data, Amazon keeps some things I've been looking at allows them to personalize what they show me. Which hopefully helps narrow down the huge raft of options I might get than just searching on dinner plates. Now, businesses can leverage technology to make better informed decisions that are actually based on signals generated by actual consumers, like me. Big data enables you to hear the voice of each consumer as opposed to consumers at large. Now, many companies, including Walmart and Target, use this information to personalize their communications with their costumers, which in turns leads to better met consumer expectations and happier customers. Which basically is to say, big data has enabled personalized marketing. Consumers are copiously generating publicly accessible data through social media sites, like Twitter or Facebook. Through such data, the companies are able to see their purchase history, what they searched for, what they watched, where they have been, and what they're interested in through their likes and shares. Let's look at some examples of how companies are putting this information to build better marketing campaigns and reach the right customers. One area we are all familiar with are the recommendation engines. These engines leverage user patterns and product features to predict best match product for enriching the user experience. If you ever shopped on Amazon, you know you get recommendations based on your purchase. Similarly, Netflix would recommend you to watch new shows based on your viewing history. Another technique that companies use is sentiment analysis, or in simple terms, analysis of the feelings around events and products. Remember the blue plates I purchased on Amazon.com? I not only can read the reviews before purchasing them, I can also write a product review once I receive my plates. This way, other customers can be informed. But more importantly, Amazon can keep a watch on the product reviews and trends for a particular product. In this case, blue plates. For example, they can judge if a product review is positive or negative. In this case, while the first review is negative, the next two reviews are positive. Since these reviews are written in English using a technique called natural language processing, and other analytical methods, Amazon can analyze the general opinion of a person or public about such a product. This is why sentiment analysis often gets referred to as opinion mining. News channels are filled with Twitter feed analysis every time an event of importance occurs, such as elections. Brands utilize sentiment analysis to understand how customers relate to their product, positively, negatively, neutral. This depends heavily on use of natural language processing. Mobile devices are ubiquitous and people almost always carry their cellphones with them. Mobile advertising is a huge market for businesses. Platforms utilize the sensors in mobile devices, such as GPS, and provide real time location based ads, offer discounts, based on this deluge of data. This time, let's imagine that I bought a new house and I happen to be in a few miles range of a Home Depot. Sending me mobile coupons about paint, shelves, and other new home related purchases would remind me of Home Depot. There's a big chance I would stop by Home Depot. Bingo! Now I would like to take a moment to analyze what kinds of big data are needed to make this happen. There's definitely the integration of my consumer information and the online and offline databases that include my recent purchases. But more importantly, the geolocation data that falls under a larger type of big data, spacial big data. We will talk about spacial data later in this class. Let's now talk about how the global consumer behavior can be used for product growth. We are now moving from personalize marketing to the consumer behavior as a whole. Every business wants to understand their consumer’s collective behavior in order to capture the ever-changing landscape. Several big data products enable this by developing models to capture user behavior and allow businesses to target the right audience for their product. Or, develop new products for uncharted territories. Let's look at this example. After an analysis of their sales for weekdays, an airline company might notice that their morning flights are always sold out, while their afternoon flights run below capacity. This company might decide to add more morning flights based on such analysis. Notice that they are not using individual consumer choices, but using all the flights purchased without consideration to who purchased them. They might, however, decide to pay closer attention to the demographic of these consumers using big data to also add similar flights in other geographical regions. With rapid advances in genome sequencing technology, the life sciences industry is experiencing an enormous draw in biomedical big data. This biomedical data is being used by many applications in research and personalized medicine. Did you know genomics data is one of the largest growing big data types? Between 100 million and 2 billion human genomes could be sequenced by year 2025. Impressive. This [INAUDIBLE] sequence data demands for between 2 exabytes and 40 exabytes in data storage. In comparison, all of YouTube only requires 1 to 2 exabytes a year. An exabyte is 10 to the power 18 bites. That is, 18 zeros after 40. Of course, analysis of such massive volumes of sequence data is expensive. It could take up to 10,000 trillion CPU hours. One of the biomedical applications that this much data is enabling is personalized medicine. Before personalized medicine, most patients without a specific type and stage of cancer received the same treatment, which worked better for some than the others. Research in this area is enabling development of methods to analyze large scale data to develop solutions that tailor to each individual, and hence hypothesize to be more effective. A person with cancer may now still receive a treatment plan that is standard, such as surgery to remove a tumor. However, the doctor may also be able to recommend some type of personalized cancer treatment. A big challenge in biomedical big data applications, like many other fields, is how we can integrate many types of data sources to gain further insight problem. In one of our future lectures, my colleagues here at the Supercomputer Center, will explain how he and his colleague have used big data from a variety of sources for personalized patient interventions. Another application of big data comes from interconnected mesh of large number of sensors implanted across smart cities. Analysis of data generated from sensors in real time allows cities to deliver better service quality to inhabitants. And reduce unwanted affect such as pollution, traffic congestion, higher than optimal cost on delivering urban services. Let's take our city, San Diego. San Diego generates a huge volumes of data from many sources. Traffic sensors, satellites, camera networks, and more. What if we could integrate and synthesize these data streams to do even more for our community? Using such big data, we can work toward making San Diego the prototype digital city. Not only for life-threatening hazards, but making our daily lives better, such as managing traffic flow more efficiently or maximizing energy savings, even as we'll see next, wildfires. If you want to read more, here's a link to the AT Kearney report, where they talk about other areas using big data. As a summary, big data has a huge potential to enable models with higher precision in many application areas. And these highly precise models are influencing and transforming business.
