Dmitry Kazakov, Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararrun Ƙwararru na Kazakhstan.
A cikin hoto: Dmitry Kazakov
Ka tuna sanannen magana cewa Big Data ya fi kama da jima'i na matasa - kowa yana magana game da shi, amma babu wanda ya san ko akwai gaske. Haka za a iya ce game da kasuwa ga bayanai kwararru (a Kazakhstan) - akwai hype, amma wanda yake a baya da shi (kuma ko akwai wani a can a duk) ba gaba daya bayyana - ba ga HR, kuma ba ga manajoji , kuma ba ga da kansu masana kimiyyar bayanai.
Mun kashe , inda suka binciki kwararru sama da 300 game da albashinsu, ayyukansu, kwarewa, kayan aikinsu da sauran su.
Mai ɓarna: Haka ne, tabbas sun wanzu, amma duk abin ba haka ba ne mai sauƙi.
Kyakkyawan fahimta. Na farko, akwai ƙarin masana kimiyyar bayanai fiye da yadda muke tsammani. Mun yi nasarar yin hira da mutane 300, a cikin su ba kawai samfura, tallace-tallace da manazarta BI ba, har ma da injiniyoyin ML da DWH, wanda ya kasance mai daɗi musamman. Ƙungiyar mafi girma ta haɗa da duk waɗanda ke kiran kansu masana kimiyyar bayanai - wannan shine 36% na masu amsawa. Yana da wuya a ce ko wannan ya shafi bukatun kasuwa ko a'a, domin kasuwar kanta kawai ake yi.
Rarraba matakan aiki yana da ruɗani - akwai kusan yawancin jagororin ƙungiyar da manajoji a matsayin ƙananan yara. Akwai dalilai da yawa na wannan. Alal misali, babban adadin ƙananan ƙungiyoyi na mutane 2-3, wanda jagoran zai iya zama ƙwararren matsakaici ko babba.
Wani dalili na iya zama hargitsi a halin yanzu yana mulki a kasuwa game da ma'auni a cikin rarraba ayyuka da ayyuka. Wani lokaci ana sanya jagororin ƙungiyar ga waɗanda kawai ke aiki shekara ɗaya ko biyu fiye da sauran, ba tare da la'akari da matakin ƙwarewa da ilimi ba. Muna ganin wannan a cikin rarraba ayyuka ta matsayi - 38% na manajoji da shugabannin ƙungiyar suna tsunduma cikin aiwatarwa da kuma wani 33% a cikin ƙididdigar ƙididdiga na asali.
Anan mun tambayi masu amsa da su tantance matakin nazari a cikin kamfanoninsu. Idan ka duba da kyau, za ka iya ganin cewa 10% na masu amsawa waɗanda ke aiki a sassan nazari na mutane 2-3 sun yi imanin cewa suna da "matakin ci gaba."
Menene "matakin ci gaba"? Tsarin BI yana aiki sosai. Akwai DWH da Big Data. Ana yin gwajin A/B akai-akai. Akwai tsarin ML da DS masu aiki a samarwa. Ana yin yanke shawara bisa bayanai kawai. Sashen sarrafa bayanai da kimiyyar bayanai na ɗaya daga cikin mahimman abubuwan da ke cikin kamfanin.
Ba shi yiwuwa a cimma duk abubuwan da ke sama tare da sashen na mutane 2-3. Ina tsammanin cewa wannan sakamakon binciken yana ɗan ƙara zafi - har yanzu samarin ba su da wanda za su kwatanta kansu da su don tantance matakin su da kyau.
Kamar yadda ake tsammani, masana kimiyyar bayanai suna ciyar da mafi yawan lokutansu ba akan babban hadadden lissafi ko injiniya ba, amma akan aiwatarwa, zazzagewa, da tsaftace bayanan. A cikin kowane ƙwarewa muna ganin preprocessing a saman 3. Amma da wuya mu ga hadaddun abubuwa kamar haɓaka ƙirar ML ko aiki tare da Big Data a saman 3 - kawai tsakanin injiniyoyin ML da DWH.
Hakanan akwai wasu fahimi na baƙin ciki. Masana sun tsara kashi 40% na ayyukansu da kansu. A Kazakhstan, ya zuwa yanzu manyan kamfanoni na unicorn ne kawai suka gwada fa'idar aiki tare da manyan bayanai kuma sun koyi yadda ake yin su da kyau. Suna watsawa ga kasuwa cewa Big Data da Machine Learning suna da kyau, kuma mataki na biyu ya biyo baya, amma ba koyaushe yana fahimtar yadda aiki tare da bayanai ke aiki ba. Sabili da haka, mun ga cewa ƙwararrun ƙwararrun sun saita ayyuka don kansu, kuma kasuwancin ba koyaushe suna san abin da suke so ba.
Na yi mamakin cewa kashi 20% na ƙwararru ba su sani ba ko kamfanin nasu yana da Data Warehouse. Ee, kuma tare da tsarin sarrafa bayanai ba komai yana da kyau sosai - 41% suna amfani da MySQL, kuma wani 34% yana amfani da PostgreSQL. Menene wannan zai iya nufi? Suna aiki maimakon tare da ƙananan bayanai.
A cikin tambaya game da tsarin ajiya, mun sake ganin MySQL har ma (!) Excel. Amma wannan na iya nuna, alal misali, cewa yawancin kamfanoni ba su da buƙatun yin aiki tare da manyan bayanai.
Anan duk abin ya sake zama shubuha. Gabaɗaya, albashi ya ɗan yi ƙasa fiye da yadda nake tsammani.
Da kaina, yana da wuya a gare ni in yi tunanin injiniyan ML wanda ke shirye ya yi aiki don tenge dubu 200 - mai yiwuwa shi ɗan ɗalibi ne. Ko dai ƙwarewar irin waɗannan ƙwararrun suna da rauni sosai, ko kuma har yanzu yana da wahala kamfanoni su iya kimanta aikin Kimiyyar Bayanai daidai. Amma watakila wannan kuma yana nuna cewa har yanzu kasuwa tana kan farkon balaga. Kuma bayan lokaci, za a kafa matakin albashi a matakin da ya dace.
source: www.habr.com