chapter1introductiondata-analyticthinking

chapter1introductiondata-analyticthinking内容摘要：

ey were interested in whether they could predict that people are expecting a baby. If they could, they would gain an advantage by making offers before their petitors. Using techniques of data science, Target analyzed historical data on customers who later were revealed to have been pregnant.  For example, pregnant mothers often change their diets, their wardrobes, their vitamin regimens, and so on. These indicators could be extracted from historical data, assembled into predictive models, and then deployed in marketing campaigns. 20 Data Science, Engineering, and DataDriven Decision Making  We will discuss predictive models in much detail as we go through the book.  For the time being, it is sufficient to understand that a predictive model abstracts away most of the plexity of the world, focusing in on particular set of indicators that correlate in some way with a quantity of interest.  Importantly, in both the Walmart and the Target example, the data analysis was not testing a simple hypothesis. Instead, the data were explored with the hope that something useful would be discovered. 21 Data Science, Engineering, and DataDriven Decision Making  Our churn example illustrates type 2 DDD problem. MegaTelCo has hundreds of millions of customers, each a candidate for defection. Ten of millions of customers have contracts expiring each month, so each one of them has an increased likelihood of defection in the near future. If we improve our ability to estimate, for a given customer, how profitable it would be for us to focus on her, we can potentially reap large benefits by applying this ability to the millions of customers in the population.  This same logic applies to many of the areas where we have seen the most application of data science and data mining: direct marketing, online advertising, credit scoring, financial trading, helpdesk management, fraud detection, search ranking, product remendation, and so on. 22 兩種由資料推動的決策 Type 1: 藉著發現數據裡所隱藏的訊息來做的決策  Walmart的例子 : 資訊長 Linda Dillman想要分析顧客購買行為，以幫忙公司預備颶風的來到 (Data: 已過颶風經過期間，顧客購買 的 資料 )  Target的例子 : 比競爭者提早預知顧客懷孕，掌握顧客可能改變消費行為的重要契機 (Data: 顧客消費資料，注意顧客是否改變飲食、衣櫥、維他命補充品等 ) 23 兩種由資料推動的決策 Type 2: 藉由分析數據將決策的精準度提升一點，但卻可以得到可觀的效益，因為這類的決策需要經常重複做，或是有廣泛的影響 電信公司預測客戶合約到期時是否會流失 : 每個月會有成千上萬的客戶合約到期，任何ㄧ位都有可能離開。 如果我們針對任ㄧ位此種客戶，能較準確的估算出如果投注行銷資源在其身上會有多少的獲利，那我們就可以套用在所有客戶而得到大的效益。 同樣的邏輯也適用其他領域的應用 : 直接行銷、線上廣告、信用評分、舞弊偵測、產品推薦等。 24 張忠謀憑直覺 獨創金字塔理論 中國時報 2020年 03月 17日  半導體產業已經進入競爭激烈的戰國時代，一個決策將會影響公司數年的發展，台積電董事長張忠謀表示，他每天都在做 決策 ，資訊當然是越多越好，但更重要的是靠決策者的直覺決定，因此他整理出了「金字塔理論」。  張忠謀的金字塔理論最底層是靠著大量的 資料 （ data）當基礎，整理出 資訊 （information）、這些資訊變成了 知識 （ Knowledge）、也提供了決策者 洞察力 （insight），進而成為創新的基石，決策者就是靠著這些尖端訊息，整合、聯想。  金字塔理論不僅可以用在公司的重要決策，判斷客戶要不要追、要不要留，也可以套用在創新、創業。 張忠謀認為創新不要狹隘，他說：「諾貝爾獎的創新，就只是一個金字塔，而創業的創新，是要把多個金字塔的頂端連結，做聯想。 」他也不斷強調， 沒有創新的創業，是不會成功的。 25 Relations between data, information, knowledge and wisdom 26 Data Science, Engineering, and DataDriven Decision Making  The diagram in figure 11 shows data science supporting datadriven decisionmaking, but also overlapping with datadriven decision making. This highlights the often overlooked fact that, increasingly, business decisions are being made automatically by puter systems. Different industries have adopted automatic decisionmaking at different rates. The finance and telemunications industries were early adopts, largely because of their precocious development of data works and implementation of massivescale puting, which allowed the aggregation and modeling of data at a large scale, as well as the application of the resultant models to decisionmaking. 27 Data Science, Engineering, and DataDriven Decision Making  In the 1990s, automated decisionmaking changed the banking and customer credit industries dramatically. In the 1990s, banks and telemunications panies also implemented massivescale systems for managing datadriven fraud control decisions.  As retail system were increasingly puterized, merchandising decisions were automated. Famous example include Harrah‘s casinos‘ reward programs and the automated remendations of Amazon and Netflix. Currently we are seeing a revolution in advertising, due in large part to a huge increase in the amount of time consumers are spending online, and the ability online to make (literally) splitsecond advertising decision (RealTime Bidding, RTB). 28 29 Netflix的例子  用戶租看的影片 75%來自系統的自動推薦  預測用戶對一部影片的評等誤差不會超過半顆星 30 Netflix的例子 (續 )  2020年 2月 Netflix推出 「 紙牌屋 」 (House of Cards)影集 • 導演：大衛芬奇 (David Fincher； 《 社群網戰 》 、 《班傑明的奇幻旅程 》 、 《 鬥陣俱樂部 》 ) • 主角：凱文史貝西 (Kevin Spacey； 《 老闆不是人 》、 《 心理醫生 》 、 《 美國心玫瑰情 》 ) • 集數：共 26 集，分兩季推出。 31 Netflix的例子 (續 ) 「紙牌屋 」 (House of Cards)影集 • 突破： 1. 只在網路上架； 2. 一次上架整季 13 集。 • 優勢：事先根據 2,900 萬 Netflix 的會員收視行為分析，根據導演、主角、政治題材的組合，找出目標觀眾，進行 精準推薦。 • 成績： IMDb 上 15,762 次評價，平均得分 ()，上架兩週後被 IMDb MOVIEmeter 評為最受歡迎的電視影集。 32 Netflix 資料分析的顧客價值主張 • 精準推薦減少客戶找尋喜好影片的時間 • 精準推薦降低客戶花時間觀看不能令其滿意影片內容的機率 • 創新影集播映方式，一季全部集數一次給足，讓客戶不用每週枯等一集 Data Processing and “Big Data”  It is important to digress here to address another point. There is a lot to data processing that is not data science—despite the impression one might get from the media. Data engineering and processing are critical to support data science, but they are more general.  For example, these days many data processing skills, systems, and technologies often are mistakenly cast as data science. To understand data science and datadriven businesses it is important to understand the differences.  Data science needs access to data and it often benefits from sophisticated data engineering that data processing technologies may facilitate, but these technologies ar。