網易傳媒數據指標體系建設實踐

{"type":"doc","content":[{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"什麼是指標體系?爲什麼建設指標體系?如何使用OSM模型和AARRR模型搭建指標體系?本文將爲大家帶來網易傳媒數據指標體系建設的實踐分享。","attrs":{}}]}],"attrs":{}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"1 什麼是指標體系?","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"百度百科的定義:指標體系是指由若干個反映社會經濟現象總體數量特徵的相對獨立又相互聯繫的統計指標所組成的有機整體。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通俗講,指標體系是將零散單點的具有相互聯繫的指標,系統化的組織起來,通過單點看全局,通過全局解決單點的問題。完整的指標體系是由指標和維度組成的。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"指標","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"指將業務單元細分後量化的度量值,它使得業務目標可描述、可度量、可拆解,它是業務和數據的結合,是統計的基礎,也是量化效果的重要依據,一般通過對某個字段的某種計算得到(比如求和、均值等)。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"維度","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"其實是指把指標按什麼角度拆分來看,這個角度用的字段就是維度(比如按照平臺,內容類型,性別等),維度可以理解爲我們看問題的角度。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"一個好的指標體系","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"能準確快速的下結論,能定義業務運作好壞的結論;","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"能滿足多數場景的歸因,即使出現異動,也能快速定位原因。","attrs":{}}]}]}],"attrs":{}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"2 爲什麼建設指標體系?","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"現階段互聯網發展已經進入精細化運營時代,精細化運營則要求產品能擁有完整、準確且有效的數據。因此爲自己的產品搭建一套數據指標體系,對於促進產品和業務增長是至關重要的。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"指標能夠量化的衡量業務的好壞,評價業務當前情況,爲業務的發展提供有效的指引,同時能使團隊成員建立共同的目標併爲之努力。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"3 如何搭建指標體系?","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"3.1 北極星指標法","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"北極星指標也叫唯一關鍵指標(OMTM,One metric that matters),產品現階段最關鍵的指標。其實簡單說來就是公司制定的發展目標,不同階段會有不同的目標。爲什麼叫“北極星”指標,其實大概的寓意就是要像北極星一樣指引公司前進的方向,目標制定最好是能符合SMART原則。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"3.2 OSM模型","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"OSM模型是三個詞縮寫:目標(Objective)、策略(Strategy)、度量(Measurement)。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"它是一套業務分析框架,並非算法模型,適用於目標已經清晰、行動方向已經明確的情況。","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/80/800ea09a484bc495762ea9d31c13b9ba.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"以網易新聞業務爲例:","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"O:","attrs":{}},{"type":"text","text":"結合北極星指標法,首先確定公司級的目標,DAU增長。接下來運行OSM模型可以把宏大的目標拆解,對應到部門內各個小組具體的、可落地、可度量的行爲上,從保證執行計劃沒有偏離大方向。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"S:","attrs":{}},{"type":"text","text":"DAU增長,可以做哪些策略呢?","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"提高新用戶規模","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"提高留存用戶規模","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"提高迴流用戶規模","attrs":{}}]}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"M:","attrs":{}},{"type":"text","text":"用什麼來評價策略是否達成?","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"新用戶數","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"留存用戶數","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"迴流用戶數","attrs":{}}]}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"這只是做了第一級拆解,我們還可以把評價的度量再做成目標,繼續拆解。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"O:","attrs":{}},{"type":"text","text":"新增用戶數","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"S:","attrs":{}},{"type":"text","text":"新用戶數增長,可以做哪些策略?比如站外渠道引流、老用戶拉新等","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"M:","attrs":{}},{"type":"text","text":"評價指標是渠道新增用戶數、老用戶拉新人數。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通過一級一級的拆解,這樣就形成了指標體系。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/98/9860856a594fc2b8a822ed66d03ff7b5.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"3.3 AARRR模型","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"麥克盧爾將創業公司最需要關注的指標分爲五大類:獲取用戶(Acquisition)、提高活躍(Activation)、提高留存率(Retention)、獲取營收(Revenue)和自傳播(Referral),簡稱AARRR。每個環節都有這個環節應該關注的指標,這些環節並不一定遵循嚴格的先後順序。","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/62/628a0449603678b23b37e199766dafaf.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"以新聞業務爲例:","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/a4/a402f0155994bbc564f5b32335e5153c.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在實際運用中,會多種模型相結合,畢竟適合的纔是最好的。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"4 網易傳媒數倉體系架構","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"網易傳媒大數據建設方法論:從業務架構設計到模型設計,從數據研發到數據服務,做到數據可管理、可追溯、可規避重複建設。杭研把這套方法論沉澱爲產品,網易有數大數據平臺(模型設計中心、指標系統、數據地圖、離線開發、自助分析、質量管理、資產管理等),方便各角色的同事使用。","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/d1/d1156aa2629825a72ae2d823d45b5b0a.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"4.1 業務線","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"傳媒業務線包括新聞、公開課、紅彩等。新聞主題域如下圖:","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/81/8136a375849c094cbb44b1a668dd53c1.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"4.2 規範定義","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"概念:","attrs":{}},{"type":"text","text":"以維度建模作爲理論基礎,構建總線矩陣,定義業務域、數據域、業務過程、度量/原子指標、維度、維度屬性、修飾詞、修飾類型、時間週期、派生指標等。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"業務域:","attrs":{}},{"type":"text","text":"比數據域更高維度的業務劃分方法,適用於特別龐大的業務系統,且業務板塊之間的指標或業務重疊性較小。例如傳媒業務可以分爲新聞、公開課、紅彩等業務板塊。新聞業務板塊包含內容、用戶、智子、PUSH、搜索、推薦等。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"業務過程:","attrs":{}},{"type":"text","text":"業務過程可以概括爲一個個不可拆分的行爲事件,如曝光,點擊,瀏覽等業務過程/事件。這裏的事件跟埋點的事件類似。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"看到這一系列的名詞,很多人可能就開始懵逼了,業務域倒還能理解,簡單來說就是對不同業務的分類;業務過程也容易理解,相當於畫業務流程圖唄。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那數據域又是何方神聖?","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數據域,是聯繫較爲緊密的數據主題的集合,是對業務對象高度概括的概念層歸類,目的是便於數據管理與應用。簡而言之,數據域就類似於我們電腦桌面要建立不同的文件夾來存儲數據,這些個文件夾名就是數據域。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"維度、維度屬性、修飾這些怎麼理解?有什麼用途?","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"維度:","attrs":{}},{"type":"text","text":"是度量的環境,用來反映業務的一類屬性,這類屬性的集合構成一個維度,可以從who-where-when-what層面來看。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"維度屬性:","attrs":{}},{"type":"text","text":"維度屬性隸屬於維度,相當於維度的具體說明,如用戶維度中性別爲男、女。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"修飾詞:","attrs":{}},{"type":"text","text":"指除了統計維度以外指標的業務場景。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"修飾類型:","attrs":{}},{"type":"text","text":"對修飾詞的抽象劃分。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"敲黑板!!!","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"underline","attrs":{}}],"text":"修飾詞和維度可以理解爲原子指標的一些限定條件,懂SQL的會更好理解一些,一般是寫SQL時,放在where語句後邊的就是修飾詞,放在group by後面的就是維度。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"指標類型:","attrs":{}},{"type":"text","text":"包含原子指標、派生指標、複合指標。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"度量/原子指標:","attrs":{}},{"type":"text","text":"原子指標和度量含義相同,某一業務行爲事件下的度量,是業務定義中不可拆分的指標,如推薦量、搜索次數。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"時間週期:","attrs":{}},{"type":"text","text":"用來明確數據統計的時間範圍或是時間點,如最近7天、自然月、截至當日等。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"派生指標","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/17/17c06fed8c790a2705890759f587bd85.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"派生指標分爲事務型指標、存量型指標","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"事務型:指標是指對業務活動進行衡量的指標,一般會對應一個事件。例如新增用戶數","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"存量型:指標是指對實體對象某些狀態的統計,例如文章總數、用戶總數,這類指標需維護原子指標及修飾詞,在此基礎上創建派生指標,對應的時間週期一般爲“歷史截至當前某時間”。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"複合指標:","attrs":{}},{"type":"text","text":"建立在原子指標、派生指標之上,通過一定運算規則形成的計算指標集合,如CTR,次均瀏覽時長等。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"4.3 模型設計","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/80/80b5c0d922ad09530797acbb8fea6c41.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(1)模型層次","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/77/77e7946c33dc93f69c323908ae62d7ef.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(2)表命名","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"ODS(ODM)層","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數據表名:odm_{業務線}_{數據源}_{庫名}_{表名}_{增量(incr)/全量(full)}_{更新時間頻次}","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"例子:","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"1)mysql:odm_{業務線}_mysql_{db_name}_{mysql_table_name}_incr_day","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"2)hbase:odm_{業務線}_hbase_{db_name}_{hbase_table_name}_incr_day","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"3)kafka:odm_{業務線}_kafka_{cluster_name}_{topic_name}_incr_day","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"4)redis:odm_{業務線}_redis_{cluster_name}_incr_day","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"5)kudu:odm_{業務線}_kudu_{db_name}_{kudu_table_name}_incr_day","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"DWD(EDM)層","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數據表名:edm_{業務域}_{一級主題}_{二級主題}_{業務過程概要}_{增量(incr)/全量(full)}_{更新時間頻次}","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"例子:","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"edm_news_content_operation_bros_detail_incr_day:新聞業務,內容主題,內容運營每日增量入池內容頭條閱讀表","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"DWS(GDM)層","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數據表名:gdm_{業務域}_{一級主題}_{二級主題}_{業務過程概要}_{增量(incr)/全量(full)}_{更新時間頻次}","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"APP(ADM)層","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數據表名:adm_{業務域}_{功能域}_{統計描述}_{更新時間頻次}","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"例子:","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"adm_news_zhizi_second_bid_ad_report_day:新聞業務,按天增量統計的智子二期競價廣告側報表","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"DIM(DDM)層","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數據表名:ddm_{業務域}_{主題}_{實體}_{增量(incr)/全量(full)}_{更新時間頻次}","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"例子:","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"ddm_content_doc_full_day:文章信息天級全量表","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"5 網易傳媒數倉指標體系建設","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"之前,業務方經常出現不同表裏的指標對不上的情況,開發查詢半天發現,兩個指標口徑是不一樣的,一個是整體的瀏覽量,一個是去掉異常數據的瀏覽量;分析師經常會問,表裏的瀏覽時長是毫秒還是秒?每每遇到這種問題,開發都要去代碼裏看看,是除了1000還是沒有;原來的指標都是寫到WIKI裏,當指標口徑發生變化時,無法及時更新,就算更新了也無法在表裏體現出來......","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"自從用了網易數帆有數大數據平臺的指標系統和模型設計中心,這些情況少多了。下面介紹一下網易傳媒數倉在大數據平臺上的應用。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"5.1 準備工作","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(1)主題域配置","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"前面我們已經根據業務線拆分了主題域,現在要把他配置到模型設計中心中,如圖:","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/db/db9cc624c680bffc9bced096ea211b4c.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(2)分層配置","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/66/6661800877c9dce59804f53841be761a.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(3)字典集配置","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/ad/ade434330a8e5d43e571682757ffb55a.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(4)表設計規則管理","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"根據我們確定的表命名規範,錄入系統","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/75/752bd0dd8d08f135247e338273be16e7.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"5.2 維度建設","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(1)查找維度","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/13/132c0285605990d1c00a7c7a16546e5e.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(2)新建維度","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/f9/f91a01b9baae3b9f71a41c3c138e9b2f.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"5.3 指標建設","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(1)查找指標","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/ff/ff024b4873de0691d2746964f309da90.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(2)新建修飾詞","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/ae/aec0b55257ab4157b34b132c1ae46631.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(3)新建指標","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/da/da6936e46d257d0eb47e53ccb00126d3.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/6d/6d91ab9239bb637195dba5ce3c49ccdf.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"5.4 表建設","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(1)表命名","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/15/157a4d3e377efc92b44a4e01fa8fdb41.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(2)基礎屬性及字段","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"字段命名","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/85/85d8ee86157ea4128bd64621b48763c8.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/5c/5c79d12142cd584eb178625ea4ca8991.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"關聯維度和指標","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/63/63ebc94b08d5935d098b68180452599f.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/39/396368ee3abf01cb07984bfeb4c0efea.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"5.5 收益","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(1)指標系統","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"方便的查詢指標間血緣關係","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/39/39f301424013d19ca0850a26df1d4c10.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"方便查詢指標被表引用情況,瞭解字段的口徑","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/40/40ae82e045a0d7cd7a750261d6c3d8e2.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"開發人員在建立指標時,一定要先確定指標的口徑,這對於後面表腳本開發非常有幫助。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"(2)模型設計中心","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"規範表的命名","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"清晰分層","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"維度一致性","attrs":{}}]}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數倉使用指標系統來管理指標的業務口徑、計算邏輯和數據來源,通過流程化的方式,建立從指標需求、指標開發、指標審覈、指標發佈的全套協作流程。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數倉使用模型設計中心按照主題域、業務過程,分層的設計方式,以維度建模作爲基本理論依據,按照維度、度量設計模型,確保模型、字段有統一的命名規範。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"6 總結","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"文章整體介紹了網易傳媒數據指標體系的建設情況,網易有數大數據平臺已經成爲數倉的主要工具,其對於維度、指標和表設計支持的很好,極大的滿足了數倉的需求。希望傳媒的大數據平臺用戶不僅是數倉,其他角色也都能參與進來。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"好啦,今天的分享就到這裏,謝謝大家。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"作者簡介","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"螺絲釘,網易傳媒數倉開發工程師。2017年入職網易,負責傳媒數倉模型設計、指標體系建設。","attrs":{}}]}]}
發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.
相關文章