The Market Sector or Sub-Space Covered

  The Market Sector or Sub-Space Covered Describe the market sector or sub-space covered in this lecture. The market sector or sub-space covered in the lecture is computational advertising. Unlike traditional advertising, where advertising agencies dominate the market, computational advertising consists of the largest advertising companies in the world, including Facebook, Instagram, Twitter, and KAYAK. Their stature as adverting companies is underpinned by their reach (to people) and their role as publishers (where people can see content). Therefore, unlike traditional agencies, the organization does not engage in traditional marketing; instead, they engage in computational advertising and are relevant for their ability to support targeted advertising. As a discipline, computational advertising spans areas of economics, computing, and machine learning. The key features in this sector include web-scaling of the audience through distributed computing, value estimation based on machine learning outcomes, targeted delivery by solving information retrieval challenges, and personalized content. Personalized content relies on machine learning, recommendation systems, and economics. The economics aspect also stands out through dynamic pricing as captured under game theory. Key organizations in the sector employ data scientists, economists, and people with computational skills.
What data science-related skills and technologies are commonly used in the sector.
Computational advertising has since surpassed television advertising in terms of reach and revenue. With an 11% growth rate, this sector will have increased opportunities for data scientists. In the United States, the growth rate is 17% annually, with the mobile sub-sector experiencing 100% annual growth. This means that the sector will increasingly require more data science related skills and technologies. Data scientist skills and technologies are vital across all the players, from advertisers, publishers, DSPs, SSPs, Ad Networks, and Exchanges to Data Aggregators. The essential skills include machine learning, data visualization, Big Data, Deep Learning, and computing. Machine learning, for instance, allows more targeted advertising, especially as advertising on mobile shapes the future of computational advertising. These skills also form part of the critical technologies utilized. In machine learning, for instance, targeted marketing requires data from historical online behavior. Thus, machine learning allows organizations to predict future behavior based on past online actions. Combining these skills and technologies with other disciplines like economics and computing forms computational advertising.
How are data and computing-related methods used in typical workflows in this sector? Illustrate with an example.
The sector relies on data and computing methods to produce advertisement products and services. Typically, for instance, whenever a person logs into Facebook, the platform already knows them through the information they provided at account creation. These include where they live, their gender, age, and preferences. Such information is combined with online behavior history obtained through data mining techniques. Thus, as soon as the feed loads after login, it provides advertisers seeking matching demographics and an opportunity to reach a potential customer. There, Facebook presents this opportunity in the action where advertisers can bid or pay for the opportunity. As a result, the winning bid is shown in the feed of the person who just logged in. Logging in with a mobile application allows for an even more targeted approach because Facebook forms part of the companies with the most reach to people, including their private information. The outcome is that products are matched with potential customer needs. Data mining and analytics are among the critical computational methods in this case. Stochastic gradient descent is among the data methods utilized where descent refers to closeness to the source of information.
What are the data science related challenges one might encounter in this domain?
There is an ar...