{"id":95283,"date":"2020-09-27T19:42:23","date_gmt":"2020-09-27T17:42:23","guid":{"rendered":"https:\/\/prohoster.info\/blog\/administrirovanie\/primenenie-low-code-v-analiticheskih-platformah"},"modified":"2020-09-27T19:42:23","modified_gmt":"2020-09-27T17:42:23","slug":"primenenie-low-code-v-analiticheskih-platformah","status":"publish","type":"post","link":"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/primenenie-low-code-v-analiticheskih-platformah","title":{"rendered":"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p>V\u00e1\u017een\u00ed \u010dten\u00e1\u0159i, p\u0159eji v\u00e1m hezk\u00fd den!<\/p>\n<p>\u00dakolem budov\u00e1n\u00ed IT platforem pro shroma\u017e\u010fov\u00e1n\u00ed a anal\u00fdzu dat se d\u0159\u00edve \u010di pozd\u011bji setk\u00e1 ka\u017ed\u00e1 spole\u010dnost, jej\u00ed\u017e obchodn\u00ed model spo\u010d\u00edv\u00e1 na intelektu\u00e1ln\u011b n\u00e1ro\u010dn\u00fdch slu\u017eb\u00e1ch nebo technicky slo\u017eit\u00fdch produktech. Vytv\u00e1\u0159en\u00ed analytick\u00fdch platforem je slo\u017eit\u00fd a \u010dasov\u011b n\u00e1ro\u010dn\u00fd \u00fakol. Ka\u017ed\u00fd \u00fakol v\u0161ak lze zjednodu\u0161it. V tomto \u010dl\u00e1nku se chci pod\u011blit o zku\u0161enosti s pou\u017eit\u00edm n\u00e1stroj\u016f low-code, kter\u00e9 pom\u00e1haj\u00ed p\u0159i vytv\u00e1\u0159en\u00ed analytick\u00fdch \u0159e\u0161en\u00ed. Tato zku\u0161enost byla z\u00edsk\u00e1na p\u0159i realizaci \u0159ady projekt\u016f zam\u011b\u0159en\u00fdch na Big Data Solutions spole\u010dnosti \u201eNeoflex\u201c. Odd\u011blen\u00ed Big Data Solutions spole\u010dnosti \u201eNeoflex\u201c se od roku 2005 zab\u00fdv\u00e1 ot\u00e1zkami v\u00fdstavby datov\u00fdch \u00falo\u017ei\u0161\u0165 a datov\u00fdch jezer, \u0159e\u0161\u00ed \u00fakoly optimalizace rychlosti zpracov\u00e1n\u00ed informac\u00ed a pracuje na metodologii \u0159\u00edzen\u00ed kvality dat.<\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/ca066793a937041577085dfb813f979e.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nVyhnout se v\u011bdom\u00e9mu shroma\u017e\u010fov\u00e1n\u00ed slab\u011b a\/nebo siln\u011b strukturovan\u00fdch dat se nikomu nepoda\u0159\u00ed. Snad i kdy\u017e p\u016fjde o mal\u00e9 podniky. P\u0159i roz\u0161i\u0159ov\u00e1n\u00ed podnik\u00e1n\u00ed se perspektivn\u00ed podnikatel setk\u00e1 s ot\u00e1zkami t\u00fdkaj\u00edc\u00edmi se v\u00fdvoje v\u011brnostn\u00edch program\u016f, chce prov\u00e9st anal\u00fdzu efektivity prodejn\u00edch m\u00edst, uva\u017euje o c\u00edlen\u00e9 reklam\u011b a tr\u00e1p\u00ed se popt\u00e1vkou po podp\u016frn\u00fdch produktech. V prvn\u00edm p\u0159ibl\u00ed\u017een\u00ed m\u016f\u017ee b\u00fdt pod\u00edl \u00fakolu vy\u0159e\u0161en \u201enarychlo\u201c. Ale s r\u016fstem podnik\u00e1n\u00ed se p\u0159echod k analytick\u00e9 platform\u011b st\u00e1v\u00e1 nevyhnuteln\u00fdm.<\/p>\n<p>Ale za jak\u00fdch okolnost\u00ed se \u00fakoly anal\u00fdzy dat mohou st\u00e1t \u00fakoly t\u0159\u00eddy \u201eRocket Science\u201c? Snad v okam\u017eiku, kdy jde o opravdu velk\u00e1 data.<br \/>\nAbychom zjednodu\u0161ili \u00fakol \u201eRocket Science\u201c, m\u016f\u017eeme sn\u00edst slona po \u010d\u00e1stech. <\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/f81e7704b91a5b95303e4240d6012730.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\n\u010c\u00edm v\u011bt\u0161\u00ed diskretnost a autonomnost budou m\u00edt va\u0161e aplikace\/slu\u017eby\/mikroslu\u017eby, t\u00edm snaz\u0161\u00ed bude pro v\u00e1s, va\u0161e kolegy a cel\u00e9 podnik\u00e1n\u00ed \u201etr\u00e1vit\u201c slona.<\/p>\n<p>K tomuto postul\u00e1tu dosp\u011bli prakticky v\u0161ichni na\u0161i klienti, kdy\u017e p\u0159eorganizovali krajinu na z\u00e1klad\u011b in\u017een\u00fdrsk\u00fdch praktik t\u00fdm\u016f DevOps.<br \/>\n<noindex><a rel=\"nofollow\" name=\"habracut\"><\/a><\/noindex><br \/>\nAle i p\u0159i \u201eodd\u011blen\u00e9, slon\u00ed\u201c diet\u011b m\u00e1me slu\u0161n\u00e9 \u0161ance na \u201ep\u0159esycen\u00ed\u201c IT krajiny. V tuto chv\u00edli byste se m\u011bli zastavit, vydechnout a pod\u00edvat se sm\u011brem k <b>low-code engineering platform<\/b>.<\/p>\n<p>Mnoh\u00e9 v\u00fdvoj\u00e1\u0159e d\u011bs\u00ed p\u0159edstava kari\u00e9rn\u00ed stagnace p\u0159i p\u0159echodu od p\u0159\u00edm\u00e9ho psan\u00ed k\u00f3du k \u201ep\u0159etahov\u00e1n\u00ed\u201c \u0161ipek v u\u017eivatelsk\u00fdch rozhran\u00edch low-code syst\u00e9m\u016f. Av\u0161ak p\u0159\u00edchod stroj\u016f nezp\u016fsobil zmizen\u00ed in\u017een\u00fdr\u016f, ale posunul jejich pr\u00e1ci na novou \u00farove\u0148!<\/p>\n<p>Poj\u010fme se pod\u00edvat, pro\u010d.<\/p>\n<p>Anal\u00fdza dat v oblasti logistiky, telekomunika\u010dn\u00edho pr\u016fmyslu, m\u00e9di\u00ed, finan\u010dn\u00edho sektoru je v\u017edy spojena s n\u00e1sleduj\u00edc\u00edmi ot\u00e1zkami: <\/p>\n<ul>\n<li>Rychlost prov\u00e1d\u011bn\u00ed automatizovan\u00e9 anal\u00fdzy;<\/li>\n<li>Mo\u017enost prov\u00e1d\u011bn\u00ed experiment\u016f bez vlivu na hlavn\u00ed tok produkce dat;<\/li>\n<li>D\u016fv\u011bryhodnost p\u0159ipraven\u00fdch dat;<\/li>\n<li>Sledov\u00e1n\u00ed zm\u011bn a verzov\u00e1n\u00ed;<\/li>\n<li>Data provenance, Data lineage, CDC;<\/li>\n<li>Rychlost dod\u00e1n\u00ed nov\u00fdch funkc\u00ed do produk\u010dn\u00edho prost\u0159ed\u00ed;<\/li>\n<li>A proslul\u00e9: n\u00e1klady na v\u00fdvoj a \u00fadr\u017ebu.<\/li>\n<\/ul>\n<p>\nTo znamen\u00e1, \u017ee in\u017een\u00fd\u0159i maj\u00ed obrovsk\u00e9 mno\u017estv\u00ed \u00fakol\u016f na vysok\u00e9 \u00farovni, kter\u00e9 lze efektivn\u011b splnit pouze t\u00edm, \u017ee se zbav\u00ed my\u0161lenek na \u00fakoly n\u00edzk\u00e9 \u00farovn\u011b v\u00fdvoje.<\/p>\n<p>P\u0159edpoklady pro p\u0159echod v\u00fdvoj\u00e1\u0159\u016f na novou \u00farove\u0148 byly evoluce a digitalizace podnik\u00e1n\u00ed. Hodnota v\u00fdvoj\u00e1\u0159e se tak\u00e9 m\u011bn\u00ed: nedostatek v\u00fdvoj\u00e1\u0159\u016f schopn\u00fdch proniknout do podstaty koncept\u016f automatizovan\u00e9ho podnik\u00e1n\u00ed je zna\u010dn\u00fd.<\/p>\n<p>Poj\u010fme si ud\u011blat analogii mezi n\u00edzko\u00farov\u0148ov\u00fdmi a vysokou \u00farovn\u00ed programovac\u00edmi jazyky. P\u0159echod od n\u00edzko\u00farov\u0148ov\u00fdch jazyk\u016f k vysokou \u00farovni \u2013 to je p\u0159echod od psan\u00ed \u201ep\u0159\u00edm\u00fdch direktiv v jazyce hardwaru\u201c k \u201edirektiv\u00e1m v jazyce lid\u00ed\u201c. To znamen\u00e1 p\u0159id\u00e1n\u00ed ur\u010dit\u00e9 vrstvy abstrakce. V tomto p\u0159\u00edpad\u011b je p\u0159echod na low-code platformy z vysokou\u00farov\u0148ov\u00fdch programovac\u00edch jazyk\u016f p\u0159echodem od \u201edirektiv v jazyce lid\u00ed\u201c k \u201edirektiv\u00e1m v jazyce podnik\u00e1n\u00ed\u201c. Pokud se najdou v\u00fdvoj\u00e1\u0159i, kter\u00e9 tento fakt znepokojuje, pak jsou asi znepokojeni u\u017e od chv\u00edle, kdy se objevil JavaScript, kter\u00fd obsahuje funkce pro t\u0159\u00edd\u011bn\u00ed pole. A tyto funkce maj\u00ed samoz\u0159ejm\u011b pod kapotou programovou implementaci jin\u00fdmi prost\u0159edky t\u00e9ho\u017e vysok\u00e9ho programov\u00e1n\u00ed.<\/p>\n<p><b>Proto je low-code jen dal\u0161\u00edm \u00farovn\u00ed abstrakce. <\/b><\/p>\n<h2>Aplika\u010dn\u00ed zku\u0161enosti s pou\u017e\u00edv\u00e1n\u00edm low-code<\/h2>\n<p>\nT\u00e9ma low-code je velmi \u0161irok\u00e9, ale nyn\u00ed bych r\u00e1d hovo\u0159il o praktick\u00e9m vyu\u017eit\u00ed \u201emalokodov\u00fdch koncept\u016f\u201c na p\u0159\u00edkladu jednoho z na\u0161ich projekt\u016f.<\/p>\n<p>Pododd\u011blen\u00ed Big Data Solutions spole\u010dnosti \u201eNeoflex\u201c se p\u0159ev\u00e1\u017en\u011b specializuje na finan\u010dn\u00ed sektor, buduje datov\u00e9 sklady a jezera dat a automatizuje r\u016fzn\u00e9 reporty. V t\u00e9to oblasti se pou\u017e\u00edv\u00e1n\u00ed low-code stalo standardem. Mezi dal\u0161\u00edmi n\u00e1stroji low-code lze zm\u00ednit prost\u0159edky pro organizaci ETL proces\u016f: Informatica Power Center, IBM Datastage, Pentaho Data Integration. Nebo Oracle Apex, kter\u00fd slou\u017e\u00ed jako prost\u0159ed\u00ed pro rychl\u00fd v\u00fdvoj rozhran\u00ed pro p\u0159\u00edstup a \u00fapravu dat. Nicm\u00e9n\u011b pou\u017eit\u00ed malokodov\u00fdch n\u00e1stroj\u016f pro v\u00fdvoj nen\u00ed v\u017edy spojeno s vytv\u00e1\u0159en\u00edm \u00fazce specializovan\u00fdch aplikac\u00ed na komer\u010dn\u00edm technologick\u00e9m stacku s jasn\u00fdm z\u00e1vislost\u00ed na dodavateli.<\/p>\n<p>S pomoc\u00ed platforem low-code je tak\u00e9 mo\u017en\u00e9 organizovat orchestraci datov\u00fdch tok\u016f, vytvo\u0159it data science platformy nebo nap\u0159\u00edklad moduly pro kontrolu kvality dat. <\/p>\n<p>Jedn\u00edm z praktick\u00fdch p\u0159\u00edklad\u016f zku\u0161enost\u00ed s vyu\u017eit\u00edm malokodov\u00fdch n\u00e1stroj\u016f je spolupr\u00e1ce \u201eNeoflex\u201c se spole\u010dnost\u00ed Mediascope, jedn\u00edm z l\u00eddr\u016f na rusk\u00e9m trhu m\u00e9di\u00ed a v\u00fdzkumu. Jedn\u00edm z \u00fakol\u016f t\u00e9to spole\u010dnosti je produkce dat, na jejich\u017e z\u00e1klad\u011b se zadavatel\u00e9 reklamy, internetov\u00e9 platformy, televizn\u00ed kan\u00e1ly, rozhlasov\u00e9 stanice, reklamn\u00ed agentury a zna\u010dky rozhoduj\u00ed o n\u00e1kupu reklamy a pl\u00e1nuj\u00ed sv\u00e9 marketingov\u00e9 komunikace. <\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/d6410c1fe93231281bdd536798fe7c66.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nMedi\u00e1ln\u00ed v\u00fdzkumy jsou technologicky n\u00e1ro\u010dnou oblast\u00ed podnik\u00e1n\u00ed. Rozpozn\u00e1v\u00e1n\u00ed videa, sb\u011br dat z za\u0159\u00edzen\u00ed analyzuj\u00edc\u00edch sledov\u00e1n\u00ed, m\u011b\u0159en\u00ed aktivity na webov\u00fdch str\u00e1nk\u00e1ch \u2013 to v\u0161e zahrnuje nutnost m\u00edt velk\u00fd IT t\u00fdm a obrovsk\u00e9 zku\u0161enosti ve v\u00fdvoji analytick\u00fdch \u0159e\u0161en\u00ed. Av\u0161ak exponenci\u00e1ln\u00ed r\u016fst mno\u017estv\u00ed informac\u00ed, po\u010dtu a rozmanitosti jejich zdroj\u016f nut\u00ed IT pr\u016fmysl dat neust\u00e1le se vyv\u00edjet. Nejjednodu\u0161\u0161\u00edm \u0159e\u0161en\u00edm pro \u0161k\u00e1lov\u00e1n\u00ed ji\u017e funguj\u00edc\u00ed analytick\u00e9 platformy Mediascope by mohlo b\u00fdt zv\u00fd\u0161en\u00ed po\u010dtu IT pracovn\u00edk\u016f. Ale mnohem efektivn\u011bj\u0161\u00edm \u0159e\u0161en\u00edm je urychlen\u00ed procesu v\u00fdvoje. Jedn\u00edm z krok\u016f, kter\u00e9 vedou t\u00edmto sm\u011brem, m\u016f\u017ee b\u00fdt pou\u017e\u00edv\u00e1n\u00ed platforem low-code.<\/p>\n<p>V dob\u011b zah\u00e1jen\u00ed projektu m\u011bla spole\u010dnost ji\u017e funk\u010dn\u00ed produktov\u00e9 \u0159e\u0161en\u00ed. Nicm\u00e9n\u011b implementace \u0159e\u0161en\u00ed na MSSQL nemohla pln\u011b splnit o\u010dek\u00e1v\u00e1n\u00ed ohledn\u011b \u0161k\u00e1lovatelnosti funkcionality p\u0159i zachov\u00e1n\u00ed p\u0159ijateln\u00e9 ceny \u00faprav.<\/p>\n<p>\u00dakol, kter\u00fd p\u0159ed n\u00e1mi st\u00e1l, byl skute\u010dn\u011b ambici\u00f3zn\u00ed \u2013 'Neoflex' a Mediascope m\u011bly vytvo\u0159it pr\u016fmyslov\u00e9 \u0159e\u0161en\u00ed za m\u00e9n\u011b ne\u017e rok, p\u0159i\u010dem\u017e MVP m\u011blo b\u00fdt uvedeno na trh ji\u017e b\u011bhem prvn\u00edho \u010dtvrtlet\u00ed od data zah\u00e1jen\u00ed prac\u00ed.<\/p>\n<p>Jako z\u00e1klad pro vybudov\u00e1n\u00ed nov\u00e9 datov\u00e9 platformy, zalo\u017een\u00e9 na low-code v\u00fdpo\u010dtech, byl vybr\u00e1n technologick\u00fd stack Hadoop. Standardem pro ukl\u00e1d\u00e1n\u00ed dat se stal HDFS s vyu\u017eit\u00edm soubor\u016f ve form\u00e1tu parquet. Pro p\u0159\u00edstup k dat\u016fm v platform\u011b byl pou\u017eit Hive, ve kter\u00e9m jsou v\u0161echny dostupn\u00e9 pohledy prezentov\u00e1ny jako extern\u00ed tabulky. Na\u010d\u00edt\u00e1n\u00ed dat do \u00falo\u017ei\u0161t\u011b bylo realizov\u00e1no pomoc\u00ed Kafka a Apache NiFi.<\/p>\n<p>Low-code n\u00e1stroj v t\u00e9to koncepci byl aplikov\u00e1n pro optimalizaci nejpracn\u011bj\u0161\u00ed \u00falohy p\u0159i vytv\u00e1\u0159en\u00ed analytick\u00e9 platformy \u2013 \u00falohy v\u00fdpo\u010dtu dat.<\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/fe62234ab0ae611a0c4a536ca68af898.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nHlavn\u00edm mechanismem pro mapov\u00e1n\u00ed dat byl vybr\u00e1n low-code n\u00e1stroj Datagram. <b>Neoflex Datagram<\/b> je prost\u0159edek pro v\u00fdvoj transformac\u00ed a tok\u016f dat. <br \/>\nPou\u017eit\u00edm tohoto n\u00e1stroje lze obej\u00edt ru\u010dn\u00ed psan\u00ed k\u00f3du ve Scal\u011b. Scala k\u00f3d je automaticky generov\u00e1n s vyu\u017eit\u00edm p\u0159\u00edstupu Model Driven Architecture.<\/p>\n<p>Z\u0159ejmou v\u00fdhodou takov\u00e9ho p\u0159\u00edstupu je zrychlen\u00ed procesu v\u00fdvoje. Av\u0161ak krom\u011b rychlosti existuj\u00ed tak\u00e9 n\u00e1sleduj\u00edc\u00ed p\u0159ednosti: <\/p>\n<ul>\n<li>N\u00e1hled obsahu a struktury zdroj\u016f\/p\u0159\u00edjemc\u016f;<\/li>\n<li>Sledov\u00e1n\u00ed p\u016fvodu objekt\u016f datov\u00e9ho toku a\u017e do jednotliv\u00fdch pol\u00ed (lineage);<\/li>\n<li>\u010c\u00e1ste\u010dn\u00e9 prov\u00e1d\u011bn\u00ed transformac\u00ed s n\u00e1hledem na mezisouhrny;<\/li>\n<li>N\u00e1hled zdrojov\u00e9ho k\u00f3du a jeho \u00faprava p\u0159ed proveden\u00edm;<\/li>\n<li>Automatick\u00e1 validace transformac\u00ed;<\/li>\n<li>Automatick\u00e9 nahr\u00e1v\u00e1n\u00ed dat 1 na 1.<\/li>\n<\/ul>\n<p>\nPrahov\u00e1 hodnota pro vstup do low-code \u0159e\u0161en\u00ed pro generov\u00e1n\u00ed transformac\u00ed nen\u00ed p\u0159\u00edli\u0161 vysok\u00e1: v\u00fdvoj\u00e1\u0159 by m\u011bl m\u00edt znalosti SQL a zku\u0161enosti s ETL n\u00e1stroji. Je v\u0161ak t\u0159eba podotknout, \u017ee gener\u00e1tory transformac\u00ed \u0159\u00edzen\u00e9 k\u00f3dem nejsou ETL n\u00e1stroji v \u0161ir\u0161\u00edm smyslu. Low-code n\u00e1stroje nemus\u00ed m\u00edt vlastn\u00ed prost\u0159ed\u00ed pro prov\u00e1d\u011bn\u00ed k\u00f3du. To znamen\u00e1, \u017ee vygenerovan\u00fd k\u00f3d se bude spou\u0161t\u011bt v tom prost\u0159ed\u00ed, kter\u00e9 bylo na clusteru k dispozici p\u0159ed instalac\u00ed low-code \u0159e\u0161en\u00ed. A to je pravd\u011bpodobn\u011b dal\u0161\u00ed plus pro low-code. Proto\u017ee vedle low-code t\u00fdmu m\u016f\u017ee pracovat \"tradi\u010dn\u00ed\" t\u00fdm, kter\u00fd implementuje funk\u010dnost nap\u0159\u00edklad v \u010dist\u00e9m Scala k\u00f3du. Za\u0159azen\u00ed \u00faprav obou t\u00fdm\u016f do v\u00fdrobn\u00edho prost\u0159ed\u00ed bude snadn\u00e9 a \"bezprobl\u00e9mov\u00e9\".<\/p>\n<p>Mo\u017en\u00e1 stoj\u00ed za to tak\u00e9 zm\u00ednit, \u017ee krom\u011b low-code existuj\u00ed i no-code \u0159e\u0161en\u00ed. A v podstat\u011b se jedn\u00e1 o r\u016fzn\u00e9 v\u011bci. Low-code v v\u011bt\u0161\u00ed m\u00ed\u0159e umo\u017e\u0148uje v\u00fdvoj\u00e1\u0159i zasahovat do generovan\u00e9ho k\u00f3du. V p\u0159\u00edpad\u011b Datagram je mo\u017en\u00e9 zobrazit a upravit generovan\u00fd k\u00f3d Scala, no-code takovou mo\u017enost nemus\u00ed poskytovat. Tento rozd\u00edl je velmi d\u016fle\u017eit\u00fd nejen z hlediska flexibility \u0159e\u0161en\u00ed, ale tak\u00e9 z pohledu komfortu a motivace v pr\u00e1ci data in\u017een\u00fdr\u016f. <\/p>\n<h2>Architektura \u0159e\u0161en\u00ed<\/h2>\n<p>\nPoj\u010fme se pokusit pochopit, jak konkr\u00e9tn\u011b low-code n\u00e1stroj pom\u00e1h\u00e1 \u0159e\u0161it \u00fakol optimalizace rychlosti v\u00fdvoje funk\u010dnosti pro v\u00fdpo\u010dty dat. Nejd\u0159\u00edve se pod\u00edvejme na funk\u010dn\u00ed architekturu syst\u00e9mu. V tomto p\u0159\u00edpad\u011b p\u0159\u00edkladem je model v\u00fdroby dat pro medi\u00e1ln\u00ed v\u00fdzkum.<\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/365f1b1cf69bf2a71854eb05c5bba809.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\n<b>Zdrojov\u00e9 datov\u00e9 informace jsou v na\u0161em p\u0159\u00edpad\u011b velmi rozmanit\u00e9 a r\u016fznorod\u00e9:<\/b><\/p>\n<ul>\n<li>Peoplemetry (TV meters) jsou programov\u011b-hardwarov\u00e1 za\u0159\u00edzen\u00ed, kter\u00e1 zaznamen\u00e1vaj\u00ed chov\u00e1n\u00ed u\u017eivatel\u016f respondent\u016f televizn\u00ed panelu \u2013 kdo, kdy a kterou televizn\u00ed stanici sledoval v dom\u00e1cnosti zapojen\u00e9 do v\u00fdzkumu. Dod\u00e1van\u00e9 informace jsou proud interval\u016f sledov\u00e1n\u00ed vys\u00edl\u00e1n\u00ed spojen\u00e9 s medi\u00e1ln\u00edm bal\u00ed\u010dkem a medi\u00e1ln\u00edm produktem. Data v procesu na\u010d\u00edt\u00e1n\u00ed do Data Lake mohou b\u00fdt obohacena demografick\u00fdmi atributy, spojen\u00edm s geografick\u00fdm um\u00edst\u011bn\u00edm, \u010dasov\u00fdmi z\u00f3nami a dal\u0161\u00edmi informacemi nezbytn\u00fdmi pro anal\u00fdzu sledovanosti ur\u010dit\u00e9ho medi\u00e1ln\u00edho produktu. Proveden\u00e1 m\u011b\u0159en\u00ed mohou b\u00fdt vyu\u017eita pro anal\u00fdzu nebo pl\u00e1nov\u00e1n\u00ed reklamn\u00edch kampan\u00ed, hodnocen\u00ed aktivity a preferenc\u00ed publika, vypracov\u00e1n\u00ed vys\u00edlac\u00edho pl\u00e1nu;<\/li>\n<li>Data mohou poch\u00e1zet z monitorovac\u00edch syst\u00e9m\u016f streamovan\u00e9ho televizn\u00edho vys\u00edl\u00e1n\u00ed a m\u011b\u0159en\u00ed sledovanosti obsahu video zdroj\u016f na internetu;<\/li>\n<li>M\u011b\u0159ic\u00ed n\u00e1stroje ve webov\u00e9m prost\u0159ed\u00ed, mezi kter\u00fdmi se nach\u00e1zej\u00ed jak site-centric, tak i user-centric po\u010ditadla. Dodavatelem dat pro Data Lake m\u016f\u017ee slou\u017eit roz\u0161\u00ed\u0159en\u00ed prohl\u00ed\u017ee\u010de research bar a mobiln\u00ed aplikace se zabudovan\u00fdm; <a class=\"wpil_keyword_link\" href=\"https:\/\/prohoster.info\/cs\/vpn\/\"   title=\"VPN\" data-wpil-keyword-link=\"linked\"  data-wpil-monitor-id=\"133\">VPN<\/a>.<\/li>\n<li>Data mohou tak\u00e9 poch\u00e1zet z platformy, kter\u00e1 konsoliduje v\u00fdsledky vypl\u0148ov\u00e1n\u00ed online dotazn\u00edk\u016f a z\u00e1v\u011bry telefonn\u00edch rozhovor\u016f v pr\u016fzkumech spole\u010dnosti;<\/li>\n<li>Dodate\u010dn\u00e9 obohacen\u00ed datov\u00e9ho jezera m\u016f\u017ee prob\u00edhat prost\u0159ednictv\u00edm na\u010d\u00edt\u00e1n\u00ed informac\u00ed z log\u016f partnersk\u00fdch spole\u010dnost\u00ed.<\/li>\n<\/ul>\n<p>\nImplementace as is na\u010d\u00edt\u00e1n\u00ed ze zdrojov\u00fdch syst\u00e9m\u016f do prim\u00e1rn\u00edho stagingu syrov\u00fdch dat m\u016f\u017ee b\u00fdt organizov\u00e1na r\u016fzn\u00fdmi zp\u016fsoby. V p\u0159\u00edpad\u011b pou\u017eit\u00ed low-code pro tyto \u00fa\u010dely je mo\u017en\u00e1 automatick\u00e1 generace skript\u016f na\u010d\u00edt\u00e1n\u00ed na z\u00e1klad\u011b metadat. V tomto p\u0159\u00edpad\u011b nen\u00ed nutn\u00e9 se dost\u00e1vat na \u00farove\u0148 v\u00fdvoje mapov\u00e1n\u00ed source to target. Pro realizaci automatick\u00e9ho na\u010d\u00edt\u00e1n\u00ed je t\u0159eba vytvo\u0159it spojen\u00ed se zdrojem, po \u010dem\u017e je nutn\u00e9 v u\u017eivatelsk\u00e9m rozhran\u00ed na\u010d\u00edt\u00e1n\u00ed definovat seznam entit, kter\u00e9 maj\u00ed b\u00fdt na\u010dteny. Vytvo\u0159en\u00ed struktury adres\u00e1\u0159\u016f v HDFS prob\u011bhne automaticky a bude odpov\u00eddat struktu\u0159e ukl\u00e1d\u00e1n\u00ed dat v syst\u00e9m\u011b zdroje.<\/p>\n<p>Nicm\u00e9n\u011b v kontextu tohoto projektu jsme se rozhodli nevyu\u017e\u00edt tuto mo\u017enost low-code platformy, proto\u017ee spole\u010dnost Mediascope ji\u017e sama zah\u00e1jila pr\u00e1ci na v\u00fdrob\u011b podobn\u00e9 slu\u017eby na z\u00e1klad\u011b Nifi + Kafka.<\/p>\n<p>Je t\u0159eba hned zd\u016fraznit, \u017ee tyto n\u00e1stroje nejsou vz\u00e1jemn\u011b zam\u011bniteln\u00e9, ale sp\u00ed\u0161e se navz\u00e1jem dopl\u0148uj\u00ed. Nifi a Kafka mohou pracovat jak v p\u0159\u00edm\u00e9m (Nifi -&gt; Kafka), tak v opa\u010dn\u00e9m (Kafka -&gt; Nifi) re\u017eimu. Pro platformu medi\u00e1ln\u00edch anal\u00fdz byl pou\u017eit prvn\u00ed zm\u00edn\u011bn\u00fd re\u017eim.<\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/8284ef5bd6d3b947998ffd42bfb66dbd.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nV na\u0161em p\u0159\u00edpad\u011b bylo nutn\u00e9, aby Nifi zpracov\u00e1val r\u016fzn\u00e9 typy dat ze syst\u00e9m\u016f-zdroj\u016f a odes\u00edlal je brokerovi Kafka. P\u0159itom bylo sm\u011brov\u00e1n\u00ed zpr\u00e1v do ur\u010dit\u00e9ho topicu Kafka prov\u00e1d\u011bno pomoc\u00ed Nifi procesor\u016f PublishKafka. Orchestraci a \u00fadr\u017ebu t\u011bchto pipeline prov\u00e1d\u00edme v vizu\u00e1ln\u00edm rozhran\u00ed. N\u00e1stroj Nifi a pou\u017eit\u00ed kombinace Nifi + Kafka lze tak\u00e9 ozna\u010dit jako low-code p\u0159\u00edstup k v\u00fdvoji, kter\u00fd m\u00e1 n\u00edzk\u00fd vstupn\u00ed pr\u00e1h do technologi\u00ed Big Data a urychluje proces v\u00fdvoje aplikac\u00ed. <\/p>\n<p>Dal\u0161\u00edm krokem v realizaci projektu bylo p\u0159enesen\u00ed do form\u00e1tu jednotn\u00e9 s\u00e9mantick\u00e9 vrstvy podrobn\u00fdch dat. V p\u0159\u00edpad\u011b, \u017ee m\u00e1 entita historick\u00e9 atributy, je v\u00fdpo\u010det prov\u00e1d\u011bn v kontextu zva\u017eovan\u00e9 partition. Pokud entita nen\u00ed historick\u00e1, m\u016f\u017ee b\u00fdt voliteln\u011b bu\u010f p\u0159epo\u010d\u00edt\u00e1no cel\u00e9 obsahu objektu, nebo se od p\u0159epo\u010dtu tohoto objektu zcela upust\u00ed (v d\u016fsledku absence zm\u011bn). V t\u00e9to f\u00e1zi doch\u00e1z\u00ed k vytv\u00e1\u0159en\u00ed kl\u00ed\u010d\u016f pro v\u0161echny entity. Kl\u00ed\u010de jsou ukl\u00e1d\u00e1ny do odpov\u00eddaj\u00edc\u00edch seznam\u016f master-objekt\u016f v Hbase, obsahuj\u00edc\u00edch shodu mezi kl\u00ed\u010di v analytick\u00e9 platform\u011b a kl\u00ed\u010di ze syst\u00e9m\u016f-zdroj\u016f. Konsolidace atom\u00e1rn\u00edch entit je doprov\u00e1zena obohacen\u00edm v\u00fdsledky p\u0159edb\u011b\u017en\u00e9ho v\u00fdpo\u010dtu analytick\u00fdch dat. Frameworkem pro v\u00fdpo\u010det dat byl Spark. Popisovan\u00e1 funkce p\u0159enosu dat do jednotn\u00e9 s\u00e9mantiky byla tak\u00e9 realizov\u00e1na na z\u00e1klad\u011b mapping\u016f low-code n\u00e1stroje Datagram.<\/p>\n<p>V c\u00edlov\u00e9 architektu\u0159e bylo nutn\u00e9 zajistit SQL p\u0159\u00edstup k dat\u016fm pro obchodn\u00ed u\u017eivatele. Pro tuto mo\u017enost byl pou\u017eit Hive. Registrace objekt\u016f v Hive se prov\u00e1d\u00ed automaticky p\u0159i aktivaci mo\u017enosti \u201eRegistr Hive Table\u201c v low-code n\u00e1stroji.<\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/a13f7260d067d3d21b7090bd8fa10aff.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<\/p>\n<h2>Spr\u00e1va toku v\u00fdpo\u010dtu<\/h2>\n<p>\nDatagram m\u00e1 rozhran\u00ed pro n\u00e1vrh workflow tok\u016f. Mapov\u00e1n\u00ed lze spou\u0161t\u011bt pomoc\u00ed pl\u00e1nova\u010de Oozie. V rozhran\u00ed pro v\u00fdvoj\u00e1\u0159e tok\u016f je mo\u017en\u00e9 vytvo\u0159it sch\u00e9mata paraleln\u00edho, sekven\u010dn\u00edho nebo na podm\u00ednk\u00e1ch z\u00e1visl\u00e9ho zpracov\u00e1n\u00ed dat. Podporov\u00e1ny jsou shell skripty a java programy. Je tak\u00e9 mo\u017en\u00e9 vyu\u017eit\u00ed <a class=\"wpil_keyword_link\" href=\"https:\/\/prohoster.info\/cs\/server\/dts-los-angeles\/\"   title=\"serveru\" data-wpil-keyword-link=\"linked\"  data-wpil-monitor-id=\"3662\">serveru<\/a> Apache Livy. Apache Livy se pou\u017e\u00edv\u00e1 ke spou\u0161t\u011bn\u00ed aplikac\u00ed p\u0159\u00edmo z v\u00fdvojov\u00e9ho prost\u0159ed\u00ed. <\/p>\n<p>Pokud m\u00e1 spole\u010dnost ji\u017e vlastn\u00edho orchestr\u00e1tora proces\u016f, je mo\u017en\u00e9 pou\u017e\u00edt REST API k integraci mapov\u00e1n\u00ed do st\u00e1vaj\u00edc\u00edho toku. Nap\u0159\u00edklad jsme m\u011bli docela \u00fasp\u011b\u0161n\u00e9 zku\u0161enosti s integrac\u00ed mapov\u00e1n\u00ed napsan\u00e9ho ve Scale do orchestr\u00e1tor\u016f napsan\u00fdch v PLSQL a Kotlin. REST API n\u00edzko-k\u00f3dov\u00e9ho n\u00e1stroje zahrnuje operace, jako je generov\u00e1n\u00ed spustiteln\u00e9ho k\u00f3du na z\u00e1klad\u011b n\u00e1vrhu mapov\u00e1n\u00ed, vol\u00e1n\u00ed mapov\u00e1n\u00ed, vol\u00e1n\u00ed sekvence mapov\u00e1n\u00ed a samoz\u0159ejm\u011b p\u0159ed\u00e1v\u00e1n\u00ed parametr\u016f v URL pro spu\u0161t\u011bn\u00ed mapov\u00e1n\u00ed.<\/p>\n<p>Krom\u011b Oozie je mo\u017en\u00e9 organizovat tok v\u00fdpo\u010dtu prost\u0159ednictv\u00edm Airflow. Nechci se p\u0159\u00edli\u0161 zab\u00fdvat porovn\u00e1v\u00e1n\u00edm Oozie a Airflow, jen \u0159eknu, \u017ee v kontextu pr\u00e1ce na projektu medi\u00e1ln\u00edch v\u00fdzkum\u016f bylo rozhodnuto ve prosp\u011bch Airflow. Hlavn\u00edmi argumenty tentokr\u00e1t byly aktivn\u011bj\u0161\u00ed komunita, kter\u00e1 vyv\u00edj\u00ed produkt, a rozvinut\u011bj\u0161\u00ed rozhran\u00ed + API.<\/p>\n<p>Airflow je tak\u00e9 v\u00fdhodn\u00fd, proto\u017ee pro popis v\u00fdpo\u010dtov\u00fdch proces\u016f se v n\u011bm pou\u017e\u00edv\u00e1 \u0161iroce obl\u00edben\u00fd Python. A v\u016fbec, platforem pro spr\u00e1vu pracovn\u00edch tok\u016f s otev\u0159en\u00fdm zdrojov\u00fdm k\u00f3dem nen\u00ed mnoho. Spu\u0161t\u011bn\u00ed a monitorov\u00e1n\u00ed prov\u00e1d\u011bn\u00ed proces\u016f (v\u010detn\u011b Ganttova diagramu) p\u0159id\u00e1v\u00e1 body do karmy Airflow. <\/p>\n<p>Form\u00e1t konfigura\u010dn\u00edho souboru pro spu\u0161t\u011bn\u00ed mapov\u00e1n\u00ed n\u00edzko-k\u00f3dov\u00e9ho \u0159e\u0161en\u00ed se stal spark-submit. K tomu do\u0161lo ze dvou d\u016fvod\u016f. Za prv\u00e9, spark-submit umo\u017e\u0148uje p\u0159\u00edmo spustit jar soubor z konzole. Za druh\u00e9, m\u016f\u017ee obsahovat v\u0161echny pot\u0159ebn\u00e9 informace pro konfiguraci pracovn\u00edho toku (co\u017e usnad\u0148uje psan\u00ed skript\u016f pro vytv\u00e1\u0159en\u00ed Dag).<br \/>\nNej\u010dast\u011bj\u0161\u00edm prvkem pracovn\u00edho toku Airflow v na\u0161em p\u0159\u00edpad\u011b se stal SparkSubmitOperator.<\/p>\n<p>SparkSubmitOperator umo\u017e\u0148uje spou\u0161t\u011bt jar soubory \u2014 zabalen\u00e9 mapov\u00e1n\u00ed Datagram s p\u0159edem definovan\u00fdmi vstupn\u00edmi parametry.<\/p>\n<p>Je t\u0159eba poznamenat, \u017ee ka\u017ed\u00e1 \u00faloha Airflow b\u011b\u017e\u00ed v samostatn\u00e9m vl\u00e1kn\u011b a nic nev\u00ed o ostatn\u00edch \u00faloh\u00e1ch. Proto se interakce mezi \u00falohami prov\u00e1d\u00ed pomoc\u00ed \u0159\u00edd\u00edc\u00edch oper\u00e1tor\u016f, jako jsou DummyOperator nebo BranchPythonOperator.<\/p>\n<p>Kombinace pou\u017eit\u00ed low-code \u0159e\u0161en\u00ed Datagram se zjednodu\u0161en\u00edm konfigura\u010dn\u00edch soubor\u016f (vytv\u00e1\u0159ej\u00edc\u00edch Dag) vedla k v\u00fdrazn\u00e9mu urychlen\u00ed a zjednodu\u0161en\u00ed procesu v\u00fdvoje datov\u00fdch tok\u016f.<\/p>\n<h2>V\u00fdpo\u010det vitr\u00edn<\/h2>\n<p>\nSnad nejinteligentn\u011bji zat\u011b\u017eovan\u00fd krok ve v\u00fdrob\u011b analytick\u00fdch dat je krok v\u00fdstavby vitr\u00edn. V kontextu jednoho z datov\u00fdch tok\u016f v\u00fdzkumn\u00e9 spole\u010dnosti v tomto kroku prob\u00edh\u00e1 p\u0159izp\u016fsoben\u00ed standardn\u00edmu vys\u00edl\u00e1n\u00ed s ohledem na \u00fapravu \u010dasov\u00fdch p\u00e1sem ve spojen\u00ed s vys\u00edlac\u00ed s\u00edt\u00ed. Tak\u00e9 je mo\u017en\u00e1 \u00faprava na m\u00edstn\u00ed vys\u00edlac\u00ed s\u00ed\u0165 (m\u00edstn\u00ed zpr\u00e1vy a reklama). Mezi jin\u00fdm se v tomto kroku prov\u00e1d\u00ed rozd\u011blen\u00ed interval\u016f nep\u0159etr\u017eit\u00e9ho sledov\u00e1n\u00ed medi\u00e1ln\u00edch produkt\u016f na z\u00e1klad\u011b anal\u00fdzy interval\u016f sledov\u00e1n\u00ed. Zde se tak\u00e9 prov\u00e1d\u00ed \u201ev\u00e1\u017een\u00ed\u201c hodnot sledov\u00e1n\u00ed na z\u00e1klad\u011b informac\u00ed o jejich v\u00fdznamnosti (v\u00fdpo\u010det korek\u010dn\u00edho koeficientu).<\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/b5f90f7d1d95c97c9819218aab9079c4.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nSamostatn\u00fdm krokem p\u0159\u00edpravy vitr\u00edn je validace dat. Algoritmus validace je spojen s aplikac\u00ed \u0159ady matematick\u00fdch v\u011bdeck\u00fdch model\u016f. Nicm\u00e9n\u011b pou\u017eit\u00ed platformy low-code umo\u017e\u0148uje rozd\u011blit slo\u017eit\u00fd algoritmus na \u0159adu jednotliv\u00fdch vizu\u00e1ln\u011b srozumiteln\u00fdch mapov\u00e1n\u00ed. Ka\u017ed\u00e9 mapov\u00e1n\u00ed vykon\u00e1v\u00e1 \u00fazk\u00fd \u00fakol. V d\u016fsledku toho je mo\u017en\u00e1 mezif\u00e1zov\u00e1 debuggov\u00e1n\u00ed, logov\u00e1n\u00ed a vizualizace f\u00e1z\u00ed p\u0159\u00edpravy dat. <\/p>\n<p>Algoritmus validace byl rozhodnut disretizovat na n\u00e1sleduj\u00edc\u00ed podkroky:<\/p>\n<ul>\n<li>Sestaven\u00ed regres\u00ed z\u00e1vislost\u00ed sledov\u00e1n\u00ed televizn\u00ed s\u00edt\u011b v regionu na sledov\u00e1n\u00ed v\u0161ech s\u00edt\u00ed v regionu za 60 dn\u00ed. <\/li>\n<li>V\u00fdpo\u010det studentizovan\u00fdch rezidu\u00ed (odchylky skute\u010dn\u00fdch hodnot od p\u0159edpov\u011bzen\u00fdch regresn\u00edmi modely) pro v\u0161echny body regrese a pro vypo\u010d\u00edtan\u00fd den.<\/li>\n<li>V\u00fdb\u011br anom\u00e1ln\u00edch p\u00e1r\u016f region-televize, kde studentizovan\u00fd zbytek vypo\u010d\u00edtan\u00e9ho dne p\u0159ekra\u010duje normu (stanovenou nastaven\u00edm operace).<\/li>\n<li>P\u0159epo\u010det upraven\u00e9ho studentizovan\u00e9ho zbytku podle anom\u00e1ln\u00edch p\u00e1r\u016f region-s\u00ed\u0165 pro ka\u017ed\u00e9ho respondenta, kter\u00fd sledoval s\u00ed\u0165 v regionu, s ur\u010den\u00edm p\u0159\u00edsp\u011bvku tohoto respondenta (velikost zm\u011bny studentizovan\u00e9ho zbytku) p\u0159i vylou\u010den\u00ed sledov\u00e1n\u00ed tohoto respondenta ze vzorku.<\/li>\n<li>Vyhled\u00e1v\u00e1n\u00ed kandid\u00e1t\u016f, jejich\u017e vylou\u010den\u00ed p\u0159iv\u00e1d\u00ed studentizovan\u00fd zbytek v\u00fdpo\u010dtov\u00e9ho dne na normu. <\/li>\n<\/ul>\n<p>\nV\u00fd\u0161e uveden\u00fd p\u0159\u00edklad potvrzuje hypot\u00e9zu, \u017ee datov\u00fd in\u017een\u00fdr m\u00e1 ji\u017e tak p\u0159\u00edli\u0161 mnoho v\u011bc\u00ed na mysli\u2026 A pokud je to skute\u010dn\u011b \u201ein\u017een\u00fdr\u201c, a ne \u201ekod\u00e9r\u201c, pak by m\u011bl strach z profesn\u00ed degradace p\u0159i pou\u017e\u00edv\u00e1n\u00ed n\u00e1stroj\u016f low-code nakonec ustoupit.<\/p>\n<h2>Co dal\u0161\u00edho m\u016f\u017ee low-code?<\/h2>\n<p>\nOblast pou\u017eit\u00ed n\u00e1stroje low-code pro d\u00e1vkovou a streamovou zpracov\u00e1n\u00ed dat bez nutnosti ru\u010dn\u00edho psan\u00ed k\u00f3du ve Scal\u011b nekon\u010d\u00ed.<\/p>\n<p>Pou\u017eit\u00ed low-code v v\u00fdvoji datov\u00fdch jezer se pro n\u00e1s stalo ji\u017e jak\u00fdmsi standardem. Asi m\u016f\u017eeme \u0159\u00edci, \u017ee \u0159e\u0161en\u00ed na stacku Hadoop opakuj\u00ed v\u00fdvojovou cestu klasick\u00fdch DWH zalo\u017een\u00fdch na RDBMS. N\u00e1stroje s n\u00edzk\u00fdm k\u00f3dem na stacku Hadoop mohou \u0159e\u0161it jak \u00fakoly zpracov\u00e1n\u00ed dat, tak \u00fakoly v\u00fdstavby kone\u010dn\u00fdch BI rozhran\u00ed. A je t\u0159eba poznamenat, \u017ee pod BI se nemus\u00ed ch\u00e1pat pouze reprezentace dat, ale i jejich editace ze strany business u\u017eivatel\u016f. Tato funkce je n\u00e1mi \u010dasto vyu\u017e\u00edv\u00e1na p\u0159i tvorb\u011b analytick\u00fdch platforem pro finan\u010dn\u00ed sektor. <\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/cde5ce9dac9fcc9356b2b0ed291e299c.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nMezi jin\u00fdmi, pomoc\u00ed low-code a zejm\u00e9na Datagram je mo\u017en\u00e9 \u0159e\u0161it \u00fakol sledov\u00e1n\u00ed p\u016fvodu objekt\u016f datov\u00e9ho proudu s atom\u00e1rnost\u00ed a\u017e na jednotliv\u00e1 pole (lineage). K tomu bylo v low-code n\u00e1stroji implementov\u00e1no propojen\u00ed s Apache Atlas a Cloudera Navigator. V podstat\u011b mus\u00ed v\u00fdvoj\u00e1\u0159 zaregistrovat sadu objekt\u016f ve slovn\u00edc\u00edch Atlas a odkazovat na registrovan\u00e9 objekty p\u0159i vytv\u00e1\u0159en\u00ed mapov\u00e1n\u00ed. Mechanismus sledov\u00e1n\u00ed p\u016fvodu dat nebo anal\u00fdza z\u00e1vislost\u00ed objekt\u016f \u0161et\u0159\u00ed spoustu \u010dasu p\u0159i pot\u0159eb\u011b prov\u00e1d\u011bt \u00fapravy v algoritmech v\u00fdpo\u010dtu. Nap\u0159\u00edklad p\u0159i sestavov\u00e1n\u00ed finan\u010dn\u00edho v\u00fdkaznictv\u00ed tato funkce umo\u017e\u0148uje l\u00e9pe zvl\u00e1dat obdob\u00ed zm\u011bn legislativy. \u010c\u00edm l\u00e9pe ch\u00e1peme meziform\u00e1tovou z\u00e1vislost v r\u00e1mci objekt\u016f podrobn\u00e9ho vrstvy, t\u00edm m\u00e9n\u011b se setk\u00e1me s \u201en\u00e1hl\u00fdmi\u201c defekty a zkr\u00e1t\u00edme po\u010det revork\u016f. <\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/0b351b116564090bd03f6adf3fc52616.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<\/p>\n<h2>Kvalita dat &amp; Low-code<\/h2>\n<p>\nDal\u0161\u00edm \u00fakolem, kter\u00fd byl realizov\u00e1n low-code n\u00e1strojem na projektu spole\u010dnosti Mediascope, byl \u00fakol t\u0159\u00eddy Kvalita dat. Zvl\u00e1\u0161tnost\u00ed implementace konvejeru kontroly dat pro projekt v\u00fdzkumn\u00e9 spole\u010dnosti byla absence vlivu na funk\u010dnost a rychlost pr\u00e1ce hlavn\u00edho toku v\u00fdpo\u010dtu dat. Pro mo\u017enost orchestraci nez\u00e1visl\u00fdmi toky kontroly dat byl pou\u017eit ji\u017e zn\u00e1m\u00fd Apache Airflow. Jakmile byla p\u0159ipravena ka\u017ed\u00e1 f\u00e1ze v\u00fdroby dat, prob\u011bhla paraleln\u011b aktivace odd\u011blen\u00e9 \u010d\u00e1sti DQ konvejeru. <\/p>\n<p>Dobrou prax\u00ed je sledovat kvalitu dat od okam\u017eiku jejich vzniku na analytick\u00e9 platform\u011b. Maj\u00edc informace o metadatech, m\u016f\u017eeme ji\u017e od okam\u017eiku, kdy informace vstupuje do prim\u00e1rn\u00ed vrstvy, kontrolovat dodr\u017eov\u00e1n\u00ed z\u00e1kladn\u00edch podm\u00ednek \u2014 not null, constraints, foreign keys. Tato funkce byla realizov\u00e1na na z\u00e1klad\u011b automaticky generovan\u00fdch mapov\u00e1n\u00ed rodiny data quality v Datagram. K\u00f3dov\u00e1 generace v tomto p\u0159\u00edpad\u011b tak\u00e9 vych\u00e1z\u00ed z metadat modelu. Na projektu spole\u010dnosti Mediascope prob\u00edhalo propojen\u00ed s metadaty produktu Enterprise Architect. <\/p>\n<p>D\u00edky propojen\u00ed low-code n\u00e1stroje a Enterprise Architect byly automaticky generov\u00e1ny n\u00e1sleduj\u00edc\u00ed kontroly:<\/p>\n<ul>\n<li>Kontrola p\u0159\u00edtomnosti hodnot \"null\" v pol\u00edch s modifik\u00e1torem \"not null\"; <\/li>\n<li>Kontrola p\u0159\u00edtomnosti duplicit prim\u00e1rn\u00edho kl\u00ed\u010de;<\/li>\n<li>Ov\u011b\u0159en\u00ed ciz\u00edho kl\u00ed\u010de entity;<\/li>\n<li>Ov\u011b\u0159en\u00ed uniqueness \u0159\u00e1dku podle sady pol\u00ed.<\/li>\n<\/ul>\n<p>\nPro slo\u017eit\u011bj\u0161\u00ed kontroly dostupnosti a p\u0159esnosti dat byl vytvo\u0159en mapping se Scala Expression, kter\u00fd p\u0159ij\u00edm\u00e1 extern\u00ed Spark SQL k\u00f3d kontroly, sestaven\u00fd analytiky v Zeppelin.<\/p>\n<p><img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/f6dcc5719077a3186d23d7bdcb412c45.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nSamoz\u0159ejm\u011b, k automatizaci kontrol je t\u0159eba p\u0159istupovat postupn\u011b. V r\u00e1mci popisovan\u00e9ho projektu tomu p\u0159edch\u00e1zely n\u00e1sleduj\u00edc\u00ed kroky:<\/p>\n<ul>\n<li>DQ, realizovan\u00e9 v notebooc\u00edch Zeppelin;<\/li>\n<li>DQ, integrovan\u00e9 do mappingu;<\/li>\n<li>DQ ve form\u011b samostatn\u00fdch velk\u00fdch mapping\u016f, obsahuj\u00edc\u00edch cel\u00fd soubor kontrol pro jednotlivou entitu;<\/li>\n<li>Univerz\u00e1ln\u00ed parametrizovan\u00e9 DQ mappingy, kter\u00e9 p\u0159ij\u00edmaj\u00ed vstupn\u00ed informace o metadatech a obchodn\u00edch kontrol\u00e1ch.<\/li>\n<\/ul>\n<p>\nHlavn\u00ed v\u00fdhodou vytvo\u0159en\u00ed slu\u017eby parametrizovan\u00fdch kontrol je zkr\u00e1cen\u00ed \u010dasu dod\u00e1n\u00ed funkcionality do produk\u010dn\u00edho prost\u0159ed\u00ed. Nov\u00e9 kontroly kvality mohou obch\u00e1zet klasick\u00fd vzor dod\u00e1n\u00ed k\u00f3du prost\u0159ednictv\u00edm v\u00fdvojov\u00fdch a testovac\u00edch prost\u0159ed\u00ed:<\/p>\n<ul>\n<li>V\u0161echny kontroly metadat jsou generov\u00e1ny automaticky p\u0159i zm\u011bn\u011b modelu v EA;<\/li>\n<li>Kontroly dostupnosti dat (ur\u010den\u00ed p\u0159\u00edtomnosti jak\u00fdchkoli dat v ur\u010dit\u00fd okam\u017eik) mohou b\u00fdt generov\u00e1ny na z\u00e1klad\u011b rejst\u0159\u00edku, kter\u00fd uchov\u00e1v\u00e1 o\u010dek\u00e1van\u00fd \u010das vzniknut\u00ed dal\u0161\u00ed d\u00e1vky dat v r\u00e1mci objekt\u016f;<\/li>\n<li>Obchodn\u00ed kontroly p\u0159esnosti dat jsou vytv\u00e1\u0159eny analytiky v notebooc\u00edch Zeppelin. Odtud jsou p\u0159\u00edmo odes\u00edl\u00e1ny do konfigura\u010dn\u00edch tabulek modulu DQ v produk\u010dn\u00edm prost\u0159ed\u00ed. <\/li>\n<\/ul>\n<p>\nRizika p\u0159\u00edm\u00e9ho dod\u00e1n\u00ed skript\u016f do produkce prakticky neexistuj\u00ed. I p\u0159i syntaktick\u00e9 chyb\u011b je maximum, co n\u00e1m hroz\u00ed, \u2013 neproveden\u00ed jedn\u00e9 kontroly, nebo\u0165 toky v\u00fdpo\u010dtu dat a tok spou\u0161t\u011bn\u00ed kontrol kvality jsou odd\u011bleny.<\/p>\n<p>V podstat\u011b, slu\u017eba DQ je trvale spu\u0161t\u011bna v produk\u010dn\u00edm prost\u0159ed\u00ed a je p\u0159ipravena za\u010d\u00edt svou pr\u00e1ci v okam\u017eiku, kdy se objev\u00ed dal\u0161\u00ed d\u00e1vka dat. <\/p>\n<h2>Na z\u00e1v\u011br<\/h2>\n<p>\nV\u00fdhoda vyu\u017eit\u00ed low-code je z\u0159ejm\u00e1. V\u00fdvoj\u00e1\u0159i nemus\u00ed vyv\u00edjet aplikaci \u201eod z\u00e1kladu\u201c. A program\u00e1tor zbaven\u00fd dodate\u010dn\u00fdch \u00fakol\u016f dosahuje rychlej\u0161\u00edch v\u00fdsledk\u016f. Rychlost na druh\u00e9 stran\u011b osvobozuje dal\u0161\u00ed \u010dasov\u00e9 zdroje k \u0159e\u0161en\u00ed ot\u00e1zek optimalizace. V tomto p\u0159\u00edpad\u011b se tedy lze spolehnout na vy\u0161\u0161\u00ed kvalitu a rychlost \u0159e\u0161en\u00ed.<\/p>\n<p><b>Samoz\u0159ejm\u011b, low-code nen\u00ed panacea a kouzlo se samo od sebe nestane:<\/b><\/p>\n<ul>\n<li>Malokodov\u00e1 odv\u011btv\u00ed proch\u00e1zej\u00ed etapou \u201ezpevn\u011bn\u00ed\u201c a dosud neexistuj\u00ed jednotn\u00e9 pr\u016fmyslov\u00e9 standardy;<\/li>\n<li>Mnoho low-code \u0159e\u0161en\u00ed nen\u00ed zdarma a jejich po\u0159\u00edzen\u00ed by m\u011blo b\u00fdt uv\u00e1\u017een\u00fdm krokem, kter\u00fd by m\u011bl b\u00fdt u\u010din\u011bn s pln\u00fdm p\u0159esv\u011bd\u010den\u00edm o finan\u010dn\u00edch v\u00fdhod\u00e1ch jejich vyu\u017eit\u00ed;<\/li>\n<li>Mnoh\u00e1 malokodov\u00e1 \u0159e\u0161en\u00ed si ne v\u017edy dob\u0159e rozum\u011bj\u00ed s GIT \/ SVN. Bu\u010f jsou nepohodln\u00e1 na pou\u017e\u00edv\u00e1n\u00ed v p\u0159\u00edpad\u011b skryt\u00ed generovan\u00e9ho k\u00f3du;<\/li>\n<li>P\u0159i roz\u0161i\u0159ov\u00e1n\u00ed architektury m\u016f\u017ee b\u00fdt pot\u0159eba \u00faprava malokodov\u00e9ho \u0159e\u0161en\u00ed \u2013 co\u017e na druh\u00e9 stran\u011b vyvol\u00e1v\u00e1 efekt \u201ep\u0159ipoutanosti a z\u00e1vislosti\u201c na poskytovateli low-code \u0159e\u0161en\u00ed.<\/li>\n<li>Po\u017eadovan\u00e1 \u00farove\u0148 zabezpe\u010den\u00ed je mo\u017en\u00e1, ale velmi pracn\u00e1 a slo\u017eit\u00e1 na realizaci motor\u016f low-code syst\u00e9m\u016f. Malokodov\u00e9 platformy by m\u011bly b\u00fdt vyb\u00edr\u00e1ny nejen na z\u00e1klad\u011b hled\u00e1n\u00ed v\u00fdhody z jejich vyu\u017eit\u00ed. P\u0159i v\u00fdb\u011bru je dobr\u00e9 si polo\u017eit ot\u00e1zku ohledn\u011b p\u0159\u00edtomnosti funkc\u00ed spr\u00e1vy p\u0159\u00edstupu a delegov\u00e1n\u00ed \/ eskalace identifika\u010dn\u00edch dat na \u00farove\u0148 cel\u00e9ho IT prost\u0159ed\u00ed organizace.<\/li>\n<\/ul>\n<p>\n<img decoding=\"async\" alt=\"Uplatn\u011bn\u00ed low-code v analytick\u00fdch platform\u00e1ch\" src=\"\/wp-content\/uploads\/2020\/09\/8ad44d2f71f519b70f5ae1e1fe722066.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nPokud v\u0161ak zn\u00e1te v\u0161echny nev\u00fdhody vybran\u00e9ho syst\u00e9mu a p\u0159esto p\u0159evl\u00e1daj\u00ed benefity z jeho pou\u017e\u00edv\u00e1n\u00ed, pak p\u0159ejd\u011bte na mal\u00fd k\u00f3d bez obav. T\u00edm sp\u00ed\u0161e, \u017ee p\u0159echod na n\u011bj je nevyhnuteln\u00fd \u2013 jako ka\u017ed\u00e1 evoluce.<\/p>\n<p>Pokud jeden v\u00fdvoj\u00e1\u0159 na low-code platform\u011b prov\u00e1d\u00ed svou pr\u00e1ci rychleji ne\u017e dva v\u00fdvoj\u00e1\u0159i bez low-code, d\u00e1v\u00e1 to spole\u010dnosti v\u00fdhodu ve v\u0161ech ohledech. Prah vstupu do low-code \u0159e\u0161en\u00ed je ni\u017e\u0161\u00ed ne\u017e u \"tradi\u010dn\u00edch\" technologi\u00ed, co\u017e m\u00e1 pozitivn\u00ed dopad na probl\u00e9m nedostatku pracovn\u00edch sil. Pou\u017eit\u00ed low-code n\u00e1stroj\u016f m\u016f\u017ee urychlit spolupr\u00e1ci mezi funk\u010dn\u00edmi t\u00fdmy a rychlej\u0161\u00ed rozhodov\u00e1n\u00ed o spr\u00e1vnosti zvolen\u00e9ho sm\u011bru datov\u011b-v\u011bdn\u00edch v\u00fdzkum\u016f. N\u00edzkok\u00f3dov\u00e9 platformy mohou b\u00fdt p\u0159\u00ed\u010dinou digit\u00e1ln\u00ed transformace organizace, proto\u017ee vyprodukovan\u00e1 \u0159e\u0161en\u00ed mohou b\u00fdt srozumiteln\u00e1 netechnick\u00fdm specialist\u016fm (zejm\u00e9na obchodn\u00edm u\u017eivatel\u016fm). <\/p>\n<p>Pokud m\u00e1te kr\u00e1tk\u00e9 term\u00edny, zat\u00ed\u017eenou obchodn\u00ed logiku, nedostatek technologick\u00fdch odborn\u00edk\u016f a pot\u0159ebujete urychlit uveden\u00ed na trh, low-code je jedn\u00edm ze zp\u016fsob\u016f, jak uspokojit va\u0161e pot\u0159eby.<\/p>\n<p>Nelze pop\u0159\u00edt v\u00fdznam tradi\u010dn\u00edch v\u00fdvojov\u00fdch n\u00e1stroj\u016f, nicm\u00e9n\u011b v mnoha p\u0159\u00edpadech je pou\u017eit\u00ed low-code \u0159e\u0161en\u00ed nejlep\u0161\u00edm zp\u016fsobem, jak zv\u00fd\u0161it \u00fa\u010dinnost \u0159e\u0161en\u00ed \u00faloh.<br \/>\n<br \/>Zdroj: <a content=\"nofollow\" rel=\"nofollow\" href=\"https:\/\/habr.com\/ru\/company\/neoflex\/blog\/520554\/\">habr.com<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>\u0423\u0432\u0430\u0436\u0430\u0435\u043c\u044b\u0435 \u0447\u0438\u0442\u0430\u0442\u0435\u043b\u0438, \u0434\u043e\u0431\u0440\u043e\u0433\u043e \u0434\u043d\u044f! \u0417\u0430\u0434\u0430\u0447\u0430 \u043f\u043e\u0441\u0442\u0440\u043e\u0435\u043d\u0438\u044f \u0418\u0422-\u043f\u043b\u0430\u0442\u0444\u043e\u0440\u043c \u0434\u043b\u044f \u043d\u0430\u043a\u043e\u043f\u043b\u0435\u043d\u0438\u044f \u0438 \u0430\u043d\u0430\u043b\u0438\u0437\u0430 \u0434\u0430\u043d\u043d\u044b\u0445 \u0440\u0430\u043d\u043e \u0438\u043b\u0438 \u043f\u043e\u0437\u0434\u043d\u043e \u0432\u043e\u0437\u043d\u0438\u043a\u0430\u0435\u0442 \u0443 \u043b\u044e\u0431\u043e\u0439 \u043a\u043e\u043c\u043f\u0430\u043d\u0438\u0438, \u0432 \u043e\u0441\u043d\u043e\u0432\u0435 \u0431\u0438\u0437\u043d\u0435\u0441\u0430 \u043a\u043e\u0442\u043e\u0440\u043e\u0439 \u043b\u0435\u0436\u0430\u0442 \u0438\u043d\u0442\u0435\u043b\u043b\u0435\u043a\u0442\u0443\u0430\u043b\u044c\u043d\u043e \u043d\u0430\u0433\u0440\u0443\u0436\u0435\u043d\u043d\u0430\u044f \u043c\u043e\u0434\u0435\u043b\u044c \u043e\u043a\u0430\u0437\u0430\u043d\u0438\u044f \u0443\u0441\u043b\u0443\u0433 \u0438\u043b\u0438 \u0441\u043e\u0437\u0434\u0430\u043d\u0438\u0435 \u0442\u0435\u0445\u043d\u0438\u0447\u0435\u0441\u043a\u0438 \u0441\u043b\u043e\u0436\u043d\u044b\u0445 \u043f\u0440\u043e\u0434\u0443\u043a\u0442\u043e\u0432. \u041f\u043e\u0441\u0442\u0440\u043e\u0435\u043d\u0438\u0435 \u0430\u043d\u0430\u043b\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0438\u0445 \u043f\u043b\u0430\u0442\u0444\u043e\u0440\u043c \u2014 \u0441\u043b\u043e\u0436\u043d\u0430\u044f \u0438 \u0442\u0440\u0443\u0434\u043e\u0437\u0430\u0442\u0440\u0430\u0442\u043d\u0430\u044f \u0437\u0430\u0434\u0430\u0447\u0430. \u041e\u0434\u043d\u0430\u043a\u043e \u043b\u044e\u0431\u0443\u044e \u0437\u0430\u0434\u0430\u0447\u0443 \u043c\u043e\u0436\u043d\u043e \u0443\u043f\u0440\u043e\u0441\u0442\u0438\u0442\u044c. \u0412 \u044d\u0442\u043e\u0439 \u0441\u0442\u0430\u0442\u044c\u0435 \u044f \u0445\u043e\u0447\u0443 \u043f\u043e\u0434\u0435\u043b\u0438\u0442\u044c\u0441\u044f \u043e\u043f\u044b\u0442\u043e\u043c \u043f\u0440\u0438\u043c\u0435\u043d\u0435\u043d\u0438\u044f [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":95284,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[688],"tags":[],"class_list":["post-95283","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-administrirovanie"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.3 - aioseo.com -->\n\t<meta name=\"description\" content=\"\u0423\u0432\u0430\u0436\u0430\u0435\u043c\u044b\u0435 \u0447\u0438\u0442\u0430\u0442\u0435\u043b\u0438, \u0434\u043e\u0431\u0440\u043e\u0433\u043e \u0434\u043d\u044f! \u0417\u0430\u0434\u0430\u0447\u0430 \u043f\u043e\u0441\u0442\u0440\u043e\u0435\u043d\u0438\u044f \u0418\u0422-\u043f\u043b\u0430\u0442\u0444\u043e\u0440\u043c \u0434\u043b\u044f \u043d\u0430\u043a\u043e\u043f\u043b\u0435\u043d\u0438\u044f \u0438 \u0430\u043d\u0430\u043b\u0438\u0437\u0430 \u0434\u0430\u043d\u043d\u044b\u0445 \u0440\u0430\u043d\u043e \u0438\u043b\u0438 \u043f\u043e\u0437\u0434\u043d\u043e \u0432\u043e\u0437\u043d\u0438\u043a\u0430\u0435\u0442 \u0443 \u043b\u044e\u0431\u043e\u0439 \u043a\u043e\u043c\u043f\u0430\u043d\u0438\u0438, \u0432 \u043e\u0441\u043d\u043e\u0432\u0435 \u0431\u0438\u0437\u043d\u0435\u0441\u0430 \u043a\u043e\u0442\u043e\u0440\u043e\u0439 \u043b\u0435\u0436\u0430\u0442 \u0438\u043d\u0442\u0435\u043b\u043b\u0435\u043a\u0442\u0443\u0430\u043b\u044c\u043d\u043e \u043d\u0430\u0433\u0440\u0443\u0436\u0435\u043d\u043d\u0430\u044f \u043c\u043e\u0434\u0435\u043b\u044c.\" \/>\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Yuri Gagarin\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/primenenie-low-code-v-analiticheskih-platformah\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.3\" \/>\n\t\t<meta property=\"og:locale\" content=\"cs_CZ\" \/>\n\t\t<meta property=\"og:site_name\" content=\"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"\ud83e\udd47\u041f\u0440\u0438\u043c\u0435\u043d\u0435\u043d\u0438\u0435 low-code \u0432 \u0430\u043d\u0430\u043b\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0438\u0445 \u043f\u043b\u0430\u0442\u0444\u043e\u0440\u043c\u0430\u0445 | ProHoster\" \/>\n\t\t<meta property=\"og:description\" content=\"\u0423\u0432\u0430\u0436\u0430\u0435\u043c\u044b\u0435 \u0447\u0438\u0442\u0430\u0442\u0435\u043b\u0438, \u0434\u043e\u0431\u0440\u043e\u0433\u043e \u0434\u043d\u044f! \u0417\u0430\u0434\u0430\u0447\u0430 \u043f\u043e\u0441\u0442\u0440\u043e\u0435\u043d\u0438\u044f \u0418\u0422-\u043f\u043b\u0430\u0442\u0444\u043e\u0440\u043c \u0434\u043b\u044f \u043d\u0430\u043a\u043e\u043f\u043b\u0435\u043d\u0438\u044f \u0438 \u0430\u043d\u0430\u043b\u0438\u0437\u0430 \u0434\u0430\u043d\u043d\u044b\u0445 \u0440\u0430\u043d\u043e \u0438\u043b\u0438 \u043f\u043e\u0437\u0434\u043d\u043e \u0432\u043e\u0437\u043d\u0438\u043a\u0430\u0435\u0442 \u0443 \u043b\u044e\u0431\u043e\u0439 \u043a\u043e\u043c\u043f\u0430\u043d\u0438\u0438, \u0432 \u043e\u0441\u043d\u043e\u0432\u0435 \u0431\u0438\u0437\u043d\u0435\u0441\u0430 \u043a\u043e\u0442\u043e\u0440\u043e\u0439 \u043b\u0435\u0436\u0430\u0442 \u0438\u043d\u0442\u0435\u043b\u043b\u0435\u043a\u0442\u0443\u0430\u043b\u044c\u043d\u043e \u043d\u0430\u0433\u0440\u0443\u0436\u0435\u043d\u043d\u0430\u044f \u043c\u043e\u0434\u0435\u043b\u044c.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/primenenie-low-code-v-analiticheskih-platformah\" \/>\n\t\t<meta property=\"og:image\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:secure_url\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:width\" content=\"350\" \/>\n\t\t<meta property=\"og:image:height\" content=\"350\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2020-09-27T17:42:23+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2020-09-27T17:42:23+00:00\" \/>\n\t\t<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"\ud83e\udd47Pou\u017eit\u00ed low-code na analytick\u00fdch platform\u00e1ch | ProHoster","description":"V\u00e1\u017een\u00ed \u010dten\u00e1\u0159i, dobr\u00fd den! \u00dakol vytvo\u0159en\u00ed IT platforem pro shroma\u017e\u010fov\u00e1n\u00ed a anal\u00fdzu dat d\u0159\u00edve \u010di pozd\u011bji vznik\u00e1 u ka\u017ed\u00e9 spole\u010dnosti, jej\u00ed\u017e podnik\u00e1n\u00ed je zalo\u017eeno na intelektu\u00e1ln\u011b n\u00e1ro\u010dn\u00e9m modelu.","canonical_url":"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/primenenie-low-code-v-analiticheskih-platformah","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"cs_CZ","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u041f\u0440\u0438\u043c\u0435\u043d\u0435\u043d\u0438\u0435 low-code \u0432 \u0430\u043d\u0430\u043b\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0438\u0445 \u043f\u043b\u0430\u0442\u0444\u043e\u0440\u043c\u0430\u0445 | ProHoster","og:description":"\u0423\u0432\u0430\u0436\u0430\u0435\u043c\u044b\u0435 \u0447\u0438\u0442\u0430\u0442\u0435\u043b\u0438, \u0434\u043e\u0431\u0440\u043e\u0433\u043e \u0434\u043d\u044f! \u0417\u0430\u0434\u0430\u0447\u0430 \u043f\u043e\u0441\u0442\u0440\u043e\u0435\u043d\u0438\u044f \u0418\u0422-\u043f\u043b\u0430\u0442\u0444\u043e\u0440\u043c \u0434\u043b\u044f \u043d\u0430\u043a\u043e\u043f\u043b\u0435\u043d\u0438\u044f \u0438 \u0430\u043d\u0430\u043b\u0438\u0437\u0430 \u0434\u0430\u043d\u043d\u044b\u0445 \u0440\u0430\u043d\u043e \u0438\u043b\u0438 \u043f\u043e\u0437\u0434\u043d\u043e \u0432\u043e\u0437\u043d\u0438\u043a\u0430\u0435\u0442 \u0443 \u043b\u044e\u0431\u043e\u0439 \u043a\u043e\u043c\u043f\u0430\u043d\u0438\u0438, \u0432 \u043e\u0441\u043d\u043e\u0432\u0435 \u0431\u0438\u0437\u043d\u0435\u0441\u0430 \u043a\u043e\u0442\u043e\u0440\u043e\u0439 \u043b\u0435\u0436\u0430\u0442 \u0438\u043d\u0442\u0435\u043b\u043b\u0435\u043a\u0442\u0443\u0430\u043b\u044c\u043d\u043e \u043d\u0430\u0433\u0440\u0443\u0436\u0435\u043d\u043d\u0430\u044f \u043c\u043e\u0434\u0435\u043b\u044c.","og:url":"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/primenenie-low-code-v-analiticheskih-platformah","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2020-09-27T17:42:23+00:00","article:modified_time":"2020-09-27T17:42:23+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"95283","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":null,"schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":null,"breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2021-02-28 11:08:24","updated":"2026-02-22 15:29:41","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/95283","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/comments?post=95283"}],"version-history":[{"count":2,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/95283\/revisions"}],"predecessor-version":[{"id":162190,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/95283\/revisions\/162190"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/media\/95284"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/media?parent=95283"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/categories?post=95283"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/tags?post=95283"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}