 Pada tahun ini, untuk mengambil kemahiran, apa yang kita harapkan untuk membuat kemahiran ini adalah untuk membantu banyak orang untuk faham roda yang berbeda dengan hubungan untuk data, dan juga kemahiran dan kisah yang anda patut mendapat untuk mencabar pertanyaan dan mencabar posisi dalam roda itu. Jadi itu akan menjadi fokus pada kemahiran ini. Terima kasih kerana menyebabkan pertanyaan yang berhati-hati kepada pointer yang saya menyebabkan. Jadi kita tidak terlalu jauh-jauh. Jangan biar saya memperkenalkan penerbangan dulu. Jadi penerbangan pertama kita adalah Eugene, yang adalah penerbangan data dari UK.ai. Jadi ruangnya adalah penerbangan data. Baiklah, ya. Saya harus pastikan semua orang tahu ruangnya adalah penerbangan data. Okay, selanjutnya kita ada Jadid. Jadid adalah dari NTUC. Jadi dia sebenarnya adalah penerbangan latihan dan penerbangan di NTUC. Selanjutnya... Jadi ruangnya untuk Jadid, tolonglah. Okay, selanjutnya kita ada Michael dari Agilent. Michael sebenarnya adalah penerbangan data dari Agilent. Jadi ruangnya untuk Jadid tolonglah. Dan selanjutnya, kita ada Wena. Wena dari RUQ 10, yang mencari pelajar di sana. Jadi sebenarnya kita sangat berjaya untuk mempunyai lebih banyak latihan dalam penerbangan. Tapi sekarang kita boleh hanya mencari satu. Tentu saja. Saya rasa kita perlu bekerja, okey? Ya. Okey, jadi kita tidak banyak lagi untuk buat. Mari kita mulakan penerbangan dulu. Jadi saya akan membuat penerbangan. Untuk saya, tolonglah. Dan apa untuk diri anda, Koo? Koo, saya penerbangan untuk hari ini. Ya. Maafkan saya, saya sebenarnya... Penerbangan Data Science & Analytics Instructor. Di sini, saya juga penerbangan co-founder... ...untuk penerbangan Data Science & Technology. Jadi hari ini kita melihat penerbangan sejarah, jadi saya menjadi penerbangan... ...dan dia membuat penerbangan. Jadi, dalam kes ini, mari kita mulakan penerbangan. Jadi, saya rasa untuk pertanyaan pertama... ...sampai... ...apa yang anda minta untuk penerbangan ini? Bolehkah anda menerbangan penerbangan anda? Apa yang anda buat hari ini? Apa yang anda buat hari ini? Apa yang anda buat hari ini? Jadi saya rasa untuk menerbangan, mungkin itu... ...Yujin? Ya. Okey, saya rasa ini adalah pertanyaan yang penting yang anda dapat... ...semasa orang lain membuat penerbangan. Saya biasanya berkata... ...saya ingin menerbangan 50% masa saya... ...menerbangan kode, mencari data... ...20% masa saya... ...mencari penerbangan... ...sehingga kita... ...pengalaman untuk mencari penerbangan... ...20% masa saya membuat... ...apa yang kita nampak... ...untuk pekerjaan... ...resource planning, pekerjaan... ...dan 10% masa saya membuat komunikasi eksternal. Seperti ini... ...berkata-kata... ...berkata-kata, berkata-kata... ...berkata-kata dengan pelanggan, pelanggan... ...saya cuba dapat... ...insentif, seorang penerbangan. Jadi saya buat hal-hal teknik atau hal-hal yang tidak. Okey. Benar? Jadi saya sendiri... ...saya paling teknik. Saya bekerja... ...tetika hari dengan pelanggan data... ...banyak kerja... ...berkata-kata... ...berkata-kata penerbangan... ...dan kemudian... ...menerbangan. Sebelum saya... ...saya penerbangan data... ...sebab semuanya... ...berkata-kata penerbangan. Jadi... ...selepas hari yang anda pergi... ...menerbangan... ...dalam pelanggan yang mereka mempunyai... ...?" ...sebaiknya anda mengarah rupanya... ...kalau anda memperkata-kata... ...untuk memperlahi... ... jet- jet keterangguhannya. Saya mengusukkannya... ...untuk mempermakasikan jenis pengalaman... ...keberkata-kata begini. Benar? Okey. Jadi kenapa hari ini... ...menjarakan pekerjaan diБ? Jadi apa yang kami buat... ...ketika kita mencuci bot... ...baiknya kami dapat menunjukkan... membuat orang yang dapat mengalami keadaan dan membuat orang yang dapat membuat keadaan membuat keadaan yang dapat membuat keadaan kita tidak mahu terdapat banyak charat dan tidak mahu membuat keadaan yang tidak membuat Jadi banyak keadaan itu membuat keadaan yang sebenarnya mengingatkan sesuatu yang tidak terdapat dan ia mempunyai untuk membaca pada hal lain, apabila hal yang berlaku, apa yang kita juga mencari adalah masalah perniagaan kadang-kadang mereka ada pertanyaan tentang A, saya akan mempunyai objek ini, jika saya mahu melihat ini apa yang berlaku di sini kita membuat perniagaan yang terdapat untuk melihat apa yang berlaku di luar sana dan melihat dan berikan keadaan yang kita membuat keadaan jadi biasanya banyak keadaan yang membuat keadaan kita akan mempunyai sesuatu keadaan untuk keadaan keadaan pada hal lain, kita juga mencari bagaimana kita juga mencari perniagaan keadaan untuk membuat keadaan yang lebih baik dan kita juga mencari pengguna yang menggunakan keadaan yang membuat mereka tahu bagaimana menggunakan keadaan yang membuat dan mereka tahu bagaimana keadaan yang menggunakan keadaan ok, jadi saya adalah pelajaran data saya bekerja di RIT Singapura jadi saya bekerja di jadi kerja saya dari 9-6 jadi sedikit fleksibel tetapi saya bekerja hanya untuk mencari kode untuk data manipulasi kemudian saya kembali saya tidak pernah dengar mereka bercakap tentang model machine dan saya mengambil data kepada model machine dan saya kembali saya mengambil keadaan semasa semua proses saya perlu kembali jadi nampaknya saya seorang pengguna yang membuat keadaan mungkin ini juga sebuah kesilapan untuk saya beritahu apa data manipulasi masyarakat kerja kerja pertama, jika anda mempunyai kerja anda, jika anda mempunyai perniagaan mereka mempunyai masalah, diperlukan masalah mereka meminta bantuan anda atau anda mempunyai idea yang terang anda boleh bawa profil atau membuat kemungkinan yang bagus untuk membuat perniagaan jika itu perniagaan perniagaan, anda boleh memulai kemudian keadaan, jika anda dapat data Datoan dari Team Data Engineer Orang yang berlaku dengan bersih dan mereka bersih untuk anda Anda hanya dapat kecepatan Bersihkan untuk menggunakan data mereka Kemudian kedua Kita cuba bersihkan data lagi Untuk memasak format Tengok model bersihkan Kemudian format yang terbaik Kemudian ke-3 adalah memasak model bersihkan Dan anda mengambil keputusan Jika anda yakin Bersihkan untuk mengambil ke-4 Dan anda perlu membuat Yang kurang baik Untuk berbicara dengan Stakeholder dari Bersihkan Dari semua perniagaan yang berlebihan Jadi Mungkin mereka perlu Memasak model bersihkan Dan anda perlu memasak mereka Memasak model bersihkan Kedua ke-4 Kedua ke-4 adalah Anda perlu melakukan beberapa kali Untuk memasak model bersihkan Untuk memasak model bersihkan Jadi Itu adalah kerja data Mungkin projek akan mengambil beberapa bulan Dan anda akan mempunyai Proses ini Terima kasih Terima kasih, saya melihat Pertanyaan pertama di atas sana Apa yang anda rasa Itu yang paling berharga untuk anda? Biar saya menambahkan Jadi Hari ini Apa yang anda menggunakan Ketua yang anda menggunakan Dan juga Adakah anda mempunyai Ada apa yang anda mempunyai Adakah anda mempunyai Kedua yang anda membantu Mungkin Kedua ke-31 Kedua yang paling berharga Memasak model bersihkan Dan ya Kedua saya Kedua saya Memasak model bersihkan Jadi apa yang saya belajar Adakah anda memasak model bersihkan Dan saya menghubungkan Kedua yang terlalu banyak dengan Linux Jadi itu mempunyai banyak Apa yang anda rasa Apa yang anda rasa Saya bermula dengan C Jawa Dan saya belajar Python Saya belajar Setelah itu saya belajar Semuanya Saya masuk ke Jawa Setelah itu, saya melepaskan Jawa, SQL Kadang-kadang kita menggunakan SCAR Live Dan ia berlaku Sebelum kita berpunyai dengan panel lain Saya akan menyebabkan banyak orang Tanya saya Jawa memang lebih mudah untuk saya Memasak beberapa pengalaman Apa yang anda fikir Jadi jika anda tahu jawa, anda dapat sangat senang dengan prinsisah dan jawa skrips. Jadi semua langgan yang saya katakan adalah dua pengalaman untuk anda. Menjelaskan atau C++ mungkin sangat mencari kerana anda perlu menjelaskan jawa. Tetapi jika perubahan tersebut, anda perlu menjelaskan jawa. Jawa juga jawa dan jawa adalah jawa yang sangat berbeda. Jawa. Jika anda tidak dapat menjelaskan jawa, jika anda perlu menjelaskan jawa, anda perlu mencari jawa untuk menggunakan jawa. Jadi mungkin saya tidak mahu berbincang dengan itu. Saya rasa jari yang anda beritahu saya. Saya rasa satu perkara yang kami lakukan adalah logis dan keperluan. Bagaimana jika seseorang berkongsi dengan masalah? Bagaimana anda boleh mencari perubahan atau perubahan di website? Apa maksud anda perubahan atau perubahan? Bagaimana perubahan hidup? Bagaimana perubahan per hari? Bagaimana perubahan perubahan perubahan? Jadi itu bukan sesuatu yang anda selalu melihat dalam kursus online. Anda akan mempunyai banyak kursus teknik untuk bekerja masyarakat dan optimisasi. Tetapi tidak ada yang mengajar anda untuk memilih kursus betul. Ya, ia adalah RMSE, tetapi pada kursus hari, kursus kerja, kursus hidup, kursus perubahan dan sebagainya. Jadi saya rasa logis dan keperluan berkongsi. Jadi itu adalah perkara yang saya rasa. Jadi, jika anda mempunyai kursus perubahan, anda memerlukan keperluan yang sangat baik untuk memahami SQL dan SQL. Kerana jika anda mempunyai kursus perubahan, seseorang lain akan mengajar anda. Dan anda memerlukan untuk memahami dengan jauh-jauh bagaimana jika anda memperkenalkan kursus perubahan. Kerana anda memperkenalkan kursus perubahan. Dan jika banyak kali anda mempunyai kursus perubahan, dan pengalaman saya yang paling besar dalam kursus perubahan, mereka selalu akan berkata bahawa kursus perubahan itu sedap dekat SQL. Jika anda mempelajari SQL, anda akan mengambil beberapa lagi kursus perubahan. Seperti jika anda memperkenalkan kursus perubahan, kursus kursus perubahan yang benar-benar seperti itu. Kursus perubahan yang itu ia ialah alasan, dan anda memastikan anda dapatkan kursus perubahan ketika anda memperkenalkan. Bagus yang lain adalah kadang-kadang anda dapat mel interrogasi yang tidak menerima, jadi ia berlaku dengan automation, Scrypting atau microsit membantu sejauh-jauh. Jadi bagaimana jika seseorang tanya, Seseorang Robbie, Ha, saya perlukan dukungan ini. Dan telah memakan merasang di bacon, kami menghidupkan ke dalamda adore itu selama selama 6 jam. Dan kami melakukannya dengan computer. Tetapi jika anda melakukannya, anda dapat beruntung. Jadi, ia adalah kebiasaan yang penting. Anda perlu menjadi sangat baik pada menjadikan data dan menjadikan data secara efekis. Jadi, ia adalah satu-satunya hal. Hal lain adalah bahawa, semasa realisasi adalah kebiasaan yang penting, ia adalah sebab sedikit desain dan komunikasi. Jadi, apa yang saya lakukan kembali dalam keadaan perniagaan, adalah yang saya meja dalam perniagaan, dan saya juga lakukan operasi. Jadi, itu membantu. Itu memang membantu kerana ia membantu anda, bagaimana anda patut berkomunikasi dengan seseorang. Yang penting adalah anda perlu faham apa yang berlaku dengan perniagaan. Jadi, jika mereka mengatakan, anda harus tidak menghubungi apa yang berlaku dengan perniagaan. Itu bukan jawapan yang berkata. Jadi, anda perlu mempunyai perniagaan yang penting untuk mengangkat pertanyaan yang berlaku. Kerana anda adalah penggunaan. Anda adalah orang yang sedang menjadikan data secara efekis. Jadi, anda sebenarnya mempunyai pertanyaan yang berlaku dengan perniagaan. Jadi, itu adalah bagaimana perniagaan yang berlaku dengan perniagaan. Biasanya anda mempunyai apa yang berlaku dengan perniagaan anda? Ya, jadi saya tidak datang dari perniagaan teknikal, tetapi perniagaan saya adalah sebenarnya dalam psikologi yang tidak mempunyai perniagaan dan perniagaan manusia kepada saya. Jadi, saya dapat tahu bahawa itu sebenarnya sangat berguna sekarang. Selain itu, perniagaan yang berlaku, saya menerima sendiri. Jadi, sangat berguna daripada itu, saya rasa. Bagus untuk menjadi peluang. Ya, sangat berguna. Jadi, saya yakin bahawa perniagaan itu adalah kemas perniagaan. Jadi, pada masa itu saya perlu melakukan beberapa perniagaan kemas perniagaan yang berlaku dengan perniagaan. Setelah itu, saya pergi ke perniagaan kemas perniagaan. Jadi, pada masa itu, saya selalu membahas beberapa perniagaan lain. Jadi, pada masa itu saya membahas perniagaan, tetapi setelah itu, saya masih belajar untuk melihat. Jadi, selama beberapa tahun. Ya, jadi ia seperti sebuah perniagaan kemas perniagaan. Untuk kerja saya, apa yang saya belajar daripada perniagaan? Jadi, apa yang digunakan daripada perniagaan yang berlaku dengan perniagaan? Ya, hanya sebuah perniagaan. Dan sebuah cara yang anda perlu berfikir bahawa perniagaan yang berlaku dengan perniagaan. Saya rasa itu berguna. Dan sekarang, sebuah perniagaan yang diberlaku. Saya perlu gunakan Python. Untuk Python yang berlaku, saya akan menyebut yang paling guna untuk diberlaku, saya menemukan penda untuk membuat perniagaan. Dan saya banyak berlaku untuk menguatkan perniagaan untuk memberitahu perniagaan yang lebih keren. Setelah itu, ia adalah pengalaman yang penting untuk perniagaan dalam model perniagaan yang diberlaku. Ia juga pilih kerana ia hanya ialah perniagaan yang diberlaku. Bagaimana jika anda mempunyai beberapa masa? Untuk mempunyai logik, Pesan Stimping, Statistik, Saya rasa komentar di sini adalah program. Semua orang sudah beritahu bahawa anda akan berprogram banyak. Saya akan tanya untuk orang lain yang mungkin ada pertanyaan, dan itu adalah, bagaimana anda semua belajar program-nya? Bagaimana mempelajari program-nya? Bagaimana saya akan mempelajari program-nya? Bagaimana çek ibu bata? Dari kechamanan perniagaan, saya akan mempelajari program-nya. Saya akan mempelajari program-nya. Bagaimana sayakan berada di sini? Tahun kecil, saya pasti sedang belajar program-nya. Jadi... Mungkin... Sekejap-kejap... Kita tak boleh berbincang sebagai programnya Tapi kita... Saya sedang berbincang dengan XLPDA Jadi itu membantu sekejap-kejap Kita membuat sedikit maklumat untuk berkata-kata Itu membantu sekejap-kejap Tetapi banyak perkara itu adalah pendidikan Untuk satu kerja saya Sebelum itu, saya sebenarnya tidak tahu Sekejap-kejap saya tidak tahu Sekejap-kejap saya tidak tahu Okey, saya belajar sekejap-kejap dalam dua minggu Ada pelajaran online dengan saya Dan setelah anda dapat membuat perjalanan Anda tahu bagaimana ia berfungsi Itu cukup untuk anda menghubungkan Dan kemudian daripada itu, saya memulangkan Pilihan R atau Python Selama dua minggu Jadi... Anda dapat mengenai bagaimana untuk mengandalkan file Bagaimana anda membuat beberapa pilihan Sehingga daripada itu, ia benar-benar baik Saya rasa belajar koding Ia sangat mudah Walaupun anda tidak mempunyai bagaimana Ada beberapa pilihan di dalam tiga minggu Python dari scratch Atau Python untuk pelajaran total Itu adalah kata-kata Hanya untuk mengikuti beberapa Mencari pilihan Atau jika pilihan itu tinggi Saya rasa ia terlalu efektif Anda hanya belajar dan mengikuti perjalanan Pertama, saya mengajar Dan hanya membuat pilihan Dan juga membuat pilihan Dan ia mudah Kerana anda tidak perlu belajar untuk saya Untuk saya Anda tidak perlu belajar semua pilihan Atau semua perkara yang anda boleh lakukan dengan Python Anda hanya perlu belajar Semua perkara yang berlainan dengan model perjalanan Jadi tidak terlalu banyak masa Untuk belajar mereka, ia adalah pilihan Saya harap semua orang dapat mencari Sangat mudah Jika hanya satu perjalanan Satu pilihan untuk saya Berskau Saya mula mengenai Perguksaan yang betul Itu sangat pengetahuan Hanya saja anda mula menulis Apa-apa pun yang anda inginkan Anda mahu lakukan dengan sain gandals Jadi saya rasa Itu satu pilihan yang pertama Dan kami sudah mempunyai Kami akan melihatnya. Jadi, sekitar dua atau tiga alasan perjalanan sepanjang minggu. Oh, okey. Saya rasa pada masa 10 atau 15 alasan, saya dapat tanggungannya. Okey. Tapi perkara yang baik adalah, apabila anda melihatnya, setelah itu, ia akan sangat mudah. Jadi, perjalanan perjalanan akan menjadi banyak bekerja. Ya. Sebenarnya, saya akan mempunyai 3 alasan sepanjang minggu. Jadi, 3 alasan perjalanan, 3 alasan perjalanan sepanjang minggu. Saya akan mengambil lebih banyak. Saya tidak tahu kreator J.Gverie yang berkata, anda harus mencuba setiap hari. Untuk sesuatu yang tidak berlaku setiap hari. Jadi, 3 alasan perjalanan, setiap hari, setiap hari, ya. Setiap hari, saya rasa lebih baik, setiap hari tidak setiap jam. Okey, bagus. Siapa yang akan melihatnya? Jadi, saya rasa bahagian pertama saya adalah R. Saya melakukan psychologi. Kita telah melakukan eksperimen. Sebaik-baiknya, bahagian programnya adalah R. Siapa yang memulakan R? Baiklah, tidak banyak lagi. Saya telah melakukan konferensi pelajaran. Saya mengambil beberapa masalah dengan R. Jadi, saya mengambil lebih banyak alasan. Kemudian, salah satu alasan saya, kita mempunyai spark. Kita akan menggunakan spark skala. Kerana alasan 5 dan API juga tidak terlalu tinggi. Jadi, kita menggunakan spark. Saya rasa saya bersyukur untuk menggunakan perjalanan kerja saya setiap hari. Jadi, perkara yang penting yang saya beritahu adalah setelah alasan 15 ini, anda dapat melaluinya. Jadi, saya rasa tidak ada jalan lain daripada semasa anda melakukannya. Ada beberapa orang yang lebih terkenal daripada orang lain. Ada beberapa orang yang belajar lebih cepat daripada orang lain. Tapi, saya rasa anda mempunyai alasan yang cukup baik. Apa yang berlainan, ia saja alasan dan alasan yang anda masukkan untuk mengambil alasan. Saya rasa ia tidak susah untuk mengambil alasan. Seperti yang Rina cakap. Okey. Jadi, mungkin saya akan menambah sesuatu. Google dan Stack Overflow adalah kawan-kawan anda. Banyak orang fikir bahawa saya perlu belajar setiap perjalanan kerja saya dan perjalanan kerja saya. Tetapi, banyak kali saya fikir anda menganggap bagaimana masalah anda. Seperti biasa, bagaimana masalah anda, bagaimana jika anda mempunyai alasan yang cukup baik, anda dapat mengambil alasan yang lebih baik daripada orang lain. Jadi, Stack Overflow dan Google. Google. Ya. Okey. Okey. Jadi, sekarang anda tahu kenapa kami mengambil alasan di Google di sini. Ya. Okey. Saya rasa untuk kami membuka alasan. Ada soalan dari alasan? Okey. Saya perlu mengambil alasan. Okey. Okey. Sebenarnya, saya mempunyai alasan ini. Jadi, mari kita lanjutkan. Saya rasa semua perjalanan akan mempunyai hari yang menarik dan hari yang menarik. Jadi, anda boleh beritahu apa yang paling menarik kerja anda dan apa yang paling menarik untuk membuat kerja anda. Saya hanya mahu memuat alasan anda. Ya. Okey. Jadi, mari kita mulakan dulu. Baiklah. Saya mulakan dari apa yang paling menarik. Kerana saya tidak mempunyai alasan komputer, jadi untuk alasan komputer, perkara yang tidak baik. Untuk menarik, saya perlu mempunyai sesuatu yang membuat alasan komputer. Alasan komputer untuk mempunyai alasan komputer saya. Jadi, itu memang yang paling menarik untuk membuat kerja anda. Dengan hari ini, saya tidak mempunyai alasan komputer. Jadi, saya rasa saya menarik alasan itu. Jadi, saya tidak dapat membuat alasan komputer dan saya tahu bahawa saya perlu menggunakan banyak alasan komputer dan alasan komputer saya. Saya merasa agak berat. Jadi, bahawa tidak membuat apa-apa yang perlu kita lakukan, tapi saya tahu anda perlu membuat sesuatu yang menarik dan menarik tapi anda tidak dapat membuat perjalanan ke sana. Segera, anda perlu mempelajari kecil untuk mempelajari. Yang paling menarik adalah apabila anda mempunyai alasan komputer dan apabila anda mempelajari alasan komputer yang baru, model alasan komputer menerima banyak alasan komputer dan anda hanya memikirkan alasan komputer yang baru, anda berasa sangat baik dengan mankind. Terima kasih. Sangat baik dengan mankind, ya? Ya. Ya. Saya hanya memikirkan alasan komputer yang baru. Okey. Terima kasih, Michael. Ya. Jadi, yang paling menarik, saya rasa, adalah kerana saya rasa bahawa terbaik alasan komputer yang besar juga berkata, setelah seketika anda datang pada hari itu, apa-apa pun, kita boleh menarik alasan ini lebih baik. Okey. Kita boleh menarik alasan kerja, anda ambil 1-5 alasan, saya ambil 6-10 alasan, anda ambil 11 alasan. Dan kemudian, anda menarik alasan dan hanya menarik alasan. Jadi, ada masa yang anda perlu menerima alasan komputer yang besar, atau anda mengatakan okey, kita perlukan alasan ini yang besar, tetapi, kita perlu pastikan alasan ini benar-benar betul. Kerana, bukan satu alasan, anda boleh salahkan itu. Jadi, anda menarik alasan, dan kemudian, anda memastikan fokus dan perlukan alasan. Tetapi, alasan yang menarik, saya rasa, adalah pada akhirnya. Jadi, kebanyakan yang menarik, iaitu, semasa alasan alasan, iaitu, apabila anda menarik, alasan masalah anda dengan, alasan masalah anda dengan, dan kemudian, anda menarik alasan. Okey, ini alasan anda, ini apa yang alasan, dan alasan yang menarik, iaitu, apabila anda beritahu, hey, kita melakukan alasan perlukan alasan, alasan dan perlukan alasan. Dan itu benar-benar benar-benar betul. Dan itu, saya rasa, menarik alasan yang lebih baik. Baiklah. Jadi, iaitu, penulis kepala itu adalah beritas, masuk ke per gibi, dan itu akan ini menjadi perlukan corrumeria masalah. Ya, okey. Baiklah. Jadi, untuk memperbaiki, kadang-kadang dia dapat menyampau. Seperti kita heelkan bahagian, kita dapat menghantar 180 mega maze van, kita hanya menghalang 1 kefullyan, tapi kemudian kita cl consultants 19 capabilities. Selepas itu, anda harus berjalan-jalan ke semua kawasan Kodwa-Kodwa dan lagi. Dan itu dapat mempunyai pembentangan dan pertanyaan dengan mereka untuk melakukannya. Tidak, tidak, tidak, tidak, tidak. Menarik adalah biasanya anda dapat belajar. Anda selalu dapat sesuatu yang baru untuk belajar. Dan itu tidak terlalu banyak yang berlaku. Semua-manyanya ada sesuatu yang baru, sesuatu yang baru, sesuatu yang baru yang baru. Jadi, anda harus menghargainya dan melihat jika anda dapat melakukan prosesnya. Bagaimana anda menggunakan itu? Apa yang benar-benar menghargai kita adalah bahawa apabila sesuatu yang saya buat, ia digunakan ke produksi. Dan kemudian, anda dapat maklumat daripada penggunaan. Hei, ini sebenarnya benar-benar benar-benar menggunakan. Dan ini bagaimana anda dapat mengubahnya lebih jauh. Saya bermaksud, ia seperti yang Michael cakap. Sebenarnya, sesuatu yang telah kita buat, adalah mengubah kisah, memperbaiki seseorang keadaan, terutamanya saya bekerja dalam keadaan keadaan. Itu sangat menarik untuk saya. Jadi, itu adalah penjelasan. Penjelasan itu bukan di mana. Model itu adalah desain. Penjelasan itu bukan di mana. Penjelasan itu di mana. Okey, mereka kata, mari kita pergi ke produksi. Dan kemudian, penjelasan kedua, seseorang sebenarnya menggunakan itu, untuk memberikan keadaan. Untuk saya, itu adalah keadaan paling menarik dari setiap projek. Dan kemudian, keadaan paling menarik adalah... keadaan keadaan. Jadi, ia sesuatu yang penting. Saya cuba buat, setelah setiap 20 hari... dari kerja sebenar, saya menghubungkan sebuah hari... untuk menghubungkan semua kisah saya... dalam keadaan keadaan, mengubah kisah saya, menulis keadaan. Tetapi, ia sangat penting. Kalau tidak, jika anda melihat projek, selama enam bulan sekarang, atau jika seseorang berjumpa dengan projek anda, cuba berkongsi dengan anda, mereka tidak tahu apa yang berlaku. Dan itu mengubah keadaan. Untuk menulis keadaan, setelah menulis keadaan, untuk lebih daripada dua hari, anda mulakan lagi... bersyukur dan anda hanya mahu menulis keadaan... yang berlaku. Yang berlaku? Ya. Baiklah. Jadi, satu perkara yang saya tahu ini... saya rasa... hal tentang... belajar. Hal tentang belajar, apabila... itu perkara yang menarik, bagaimanapun, jika ada aloritam baru yang keluar, aloritam yang menarik, dan sebagainya. Itu yang menarik. Semua orang kami sangat gembira. Sekarang, mari kita bercakap lebih tentang belajar. Kita bercakap tentang bagaimana kita... Jadi, anda boleh berkongsi lebih banyak... bagaimana anda... belajar, bahawa teknologi... teknologi yang kita belajar dengan... mengubah keadaan dengan cepat. Dan... sentiasa menerima keadaan sejauh. Bagaimana anda belajar? Bagaimana anda menerima keadaan? Bagaimana anda menerima keadaan? Bagaimana anda menerima keadaan? Bagaimana anda menerima keadaan? Bagaimana anda menerima keadaan? Saya akan beritahu dua cara. Dua cara. Pertama, anda perlu belajar untuk menghidupkan keadaan. Apa yang terbaik untuk digaruhkan? Saya rasa yang terbaik untuk digaruhkan... adalah untuk belajar sesuatu yang anda boleh menghidupkan dengan cepat. Sehingga anda dapat melihat keadaan, anda dapat dapatkan keadaan dengan cepat. Tidak ada yang menerima keadaan dengan cepat. Jika anda menggunakan pihologi. Siapa? Tidak. Saya bermaksud... Bagaimana keadaan? Bukan cara membuat perjalanan. Bukan cara membuat perjalanan. Bukan cara membuat perjalanan. Saya rasa... Jika anda ingin berada di luar ini. Anda hanya perlu membuat sebuah masa untuk membuat perjalanan. Perjalanan bergerak dengan cepat. Pemajar kata-kata itu membuat perjalanan. Pemajar kata-kata di ganan setiap masa. Pemajar kata-kata ini membuat perjalanan. Kemudian ada masa untuk menerima kata-kata. Memahatkan apa yang orang lakukan, etc. Jadi saya cuba membaca... 2 kata-kata setiap bulan. Dalam saya, kami membuat perjalanan setiap bulan. Setiap bulan, semua orang berkongsi kata-kata. Jadi itu mungkin kita... menerima keadaan. Menerima kata-kata, etc. Jadi itu bagaimana anda perlu... membuat keadaan dengan teknologi. Tapi satu perkara yang sebenarnya... kemudian ada juga... keadaan. Ada beberapa kata-kata yang berkongsi... di sebrang semuanya yang anda belajar. Untuk contoh kata-kata. Bagaimana anda membuat kata-kata yang berkongsi? Jadi yang lain orang dapat... mudah membuat keadaan. Ini menggantikan... segalanya. Bagaimana anda membuat perjalanan? Anda membeli banyak masa untuk mencari data? Apa yang paling cepat untuk membuat perjalanan? Ada orang yang boleh melakukan dalam 4 bulan. Ada orang yang boleh melakukan dalam 4 tahun. Jadi... Itu adalah sebuah skill yang berkongsi... segala-galanya yang anda lakukan. Bagaimana anda membuat dokumentasi? Bagaimana anda menjelaskan pukul untuk... menjadi kode-kode bagi... faham 2.4 itu... ada ialah sebuah bulan... atau anda tidak tahu apa yang berlaku di sepanjang pilihan... Bagaimana anda menjelaskan pukul untuk... menjadi bagi jika anda memahami apa yang berlaku... Bagaimana anda memulai Google? Bagaimana anda memahamkan sebuah pukul? Bagaimana anda memulai metas... segala-galanya? Kita harus mencepati metas di atas... ...pengalaman teknikal yang berlainan sepanjang bulan. Terima kasih. Saya akan kata, saya rasa anda harus belajar apa yang anda suka... ...dan jangan mengambil aktiviti rumah yang anda perlu lakukan... ...saya perlu belajar sesuatu yang penting. Dan ada banyak perkara baru setiap hari... ...tidak ada cara anda dapat mencari banyak masa untuk belajar itu. Jadi, saya rasa saya akan menolakkan kaitan saya... ...untuk apa yang saya mahu belajar. Saya juga suka belajar perkara yang penting... ...yang berlainan sepanjang bulan. Biasanya saya akan belajar Python atau Java Core untuk belajar itu... ...tapi biasanya saya akan belajar perkara yang penting. Jadi, belajar perkara yang penting seperti ini... ...saya juga akan menyebabkan perkara yang penting. Okey, pada masa itu... ...saya akan menyebabkan perkara itu. Apakah perkara yang berlainan sepanjang bulan? Apakah perkara yang berlainan... ...saya akan menyebabkan perkara yang berlainan sepanjang bulan? Saya rasa setiap perkara yang penting. Setiap perkara yang penting? Ya, jadi... ...saya akan mencari perkara yang penting hari ini... ...untuk perkara yang penting hari ini. Jadi, bagaimana anda mencari... ...pembulan yang kamu tidak dapat mencari? Ini adalah masalah yang terjadi. Jadi, jika kamu cuba mencari algodam yang berlainan... ...atau 70 atau even 11 itu... ...saya juga akan mencari perkara yang penting... ...kalau kamu mencari perkara yang penting. Jadi, kita akan mencari perkara yang penting. Memang, apakah perkara yang penting... ...tergantung pada begini? Saya tidak fikir perkara yang berlainan yang baru... ...sebelum itu. Jadi, apa yang terkenal yang paling terkenal yang anda pergi ke? Saya pergi ke program. Program? Okey. Program apa? Okey, ya. Saya pergi. Terima kasih. Terima kasih. Pada tempat pertama, anda dapat belajar dan melepaskan kumpulan Facebook. Saya berada di Singapura. Sudah tentu, jika anda melihat itu, kami juga mencari keadaan keadaan keadaan keadaan keadaan keadaan keadaan keadaan keadaan. Sebenarnya, sebaik sahaja, seperti yang anda menelefon dan yang anda menelefon, banyak jadi yang anda belajar dengan masalah keadaan dengan masalah tersebut. Hari ini, currently, perkara Pournayah agak perlahan-lahan. Saya ingin memberihangkan sesuatu lain. Kemudian anda melihat keadaan keadaan keadaan keadaan keadaan keadaan keadaan, mengembangkan jalan-jalan, bercukaan, berbincangkan dengan orang dan mencari perjalanan keadaan keadaan. Kerana kadang-kadang tiada orang yang paling baik dengan pengedaran. Selepas anda belajar sesuatu, seseorang akan mengenai bahaya. Ini perkara yang besar yang keisteri, adalah perkara yang terlalu besar? Jadi, anda memang perlu melihat. Jadi, itu adalah banyak perkara yang terlalu besar. Sebenarnya, anda perlu belajar bagaimana memasukkan kebebasan anda dan terutamanya di sekeliling. Jika anda mempunyai pertanyaan yang unik, dan anda fikir ia sangat unik, bagaimana anda akan memasukkan yang tidak akan terlalu baik? Ini bukan pertanyaan yang terlalu besar. Kerana saya telah melakukannya beberapa kali sebelumnya. Saya fikir saya mempunyai pertanyaan yang baik, dan jika anda akan memasukkan kebebasan anda, hanya dengan berlaku dengan kegembaraan Anda. Kerana anda akan memasukkan kebebasan anda dan berlaku dengan kegembaraan Anda. Kita akan bertanya kepada anda. Ini adalah perkara yang penting daripada kamu, aka sengaja perkara yang terlalu besar. Bersiap-sebelah. Untuk tetap kocok tentang perkara yang penting, saya akan beri menarik untuk banyak perkara yang penting dan menerima peruteriaan. You can get some updates, or you can just get new machine learning updates from the news reports. Otherwise, I think most of my updates are from my working environment. For example, my colleagues are quite active in the blogging or in the K-POP report series. And they will just pass those new information to me before I really get it myself. Also, the head of RIT, so I put it into the technology. He was really active about all of those new information, new models, new technology. He really want to bring those new things to RIT. So at any time we have all hands-on, he will also report to us. Okay, I think continuing on the learning part. I saw one question that I think I would like to ask also. And that is, I think some of us here have taken a masters. And I know a few of the panelists have already taken a masters, or currently taking masters or have taken masters as well. So maybe can you share a bit more how the masters have helped in your work. All the masters that have helped in your work. Or if you have taken a book camp, how that book camp has really helped in your work. Michael, what's up? Masters program. So I did a masters in SMU, MITB, analytics program masters. So first of all, you need to get into the masters program knowing what you want out of it. You have to have a goal before you go in. It's not a magical upgrade. It's like you don't just go to a cause and then suddenly you're much better person. A lot of it is that you can get in touch with the academics, the professors. The way I found that I got the most value out of it is that you basically have to experiment and read up on your own a lot. Then you go to the class and test it and say that I tried this. Is this going to work? Or I have this new approach. So there are a few times where I actually said there's some stuff that the professors went, what's this new approach? And actually it works out quite well. And that's where you learn out of the program, I think. It's a lot of it. It's also the people you get around with because mine is physical analog, so-called analog masters that you meet. So it's a lot of classmates, interact with people from different areas you get that experience because we have guys who have worked in banking in my 20 years or 20 years. And their experience actually is very valuable. So that's, I think, one big benefit of the masters program is that you get to be in touch with a lot of different people from different areas with varying amount of experience. And of course you get in touch with the academic science program. I know Michael. So I know Michael actually took a dive as in like he designed and then took a masters full-time. So after that, after the masters he came and looked up for a job and all this, can you share a bit more experience of using the masters and looking up for jobs? What are the things you look up for maybe those who are taking a masters now? So first of all, having a masters is just one more question that you will ask. So what are you doing in your masters? What really, really helps you is that during a masters program or whichever learning program you have, start creating a portfolio of your projects if you can put them up. Because eventually the current company I got hired at actually the person said yeah, the CV looks okay. Then he click on the link on the profile on the portfolio that I had and he said, yeah, this looks really interesting and this looks like what we want to do. So I guess truth that if you take a deep dive and say I want to take a masters so that I jump from one road to another road because before this I was in marker research. But it really helps a lot if you get a portfolio up make sure you run through enough projects that exposes you to different things not just what you're comfortable with. So in fact I actually took some backing models that has nothing to do with it but I realised that oh, this actually helps a lot. In fact, when I first started my masters program I was like okay, okay, I'm not going to be very fancy for it. But now I have, yeah, this really okay, you never know what you want me in the future I think it's good to keep a wide graph on your masters program. So I did masters in computer science if you are just working as a programmer or engineer I don't really think there's a lot of difference between a programmer who has an undergraduate or a master's but the main thing I learnt was I did like masters per research we got to write paper system. So, I think which I learnt was like when you do write papers you have to spend a lot of time figuring out has this been done before and which all labs or what approach did they use you have to decide the plan but that skill has been very ridah chinese academic data chinese academic paper to polish up my training so, I agree it's a good chance to look for all the academic stuff okay, so Jay, did you took your masters full time or full time so how was it like to then after that go back to the industry again yeah, so it's a big story so let's take a poc one first I did masters per research I almost went into PhD I even applied to few PhD programs I luckily got a job in Singapore in a startup so, I thought maybe I should earn some money now and think about PhD later and that's how I got but as soon as I was back in industry I really like the job so, after that I did it okay usually okay, so my background was in psychology so as I was working with using data I felt that there was a sort of a ceiling I was lacking some of the fundamentals in computer science so therefore I decided to take a masters in computer science I think what Michael and Jenny mention, the component about writing in masters or at least I had to replicate some of the for example, the deep mind experiments on reinforcement learning or I had to build some very core fundamental algorithms from scratch like alpha-peta pruning I think that gives you a lot of practice, a lot of rigor and sometimes you might actually, almost all the time I had to write up about my report my findings it really holds your ability to write so I think in the masters you get very structured practice get very structured feedback on terms of your programming and your writing I think the other thing is also you learn things and you can apply it immediately at work hopefully almost everything you learn you can almost find a way to apply it so I think it's useful okay, so actually you didn't took a part-time masters so for you since you're working and taking masters at the same time what is the key essential ingredients to help you get by day to day since you got to manage both I think the key ingredient is that I basically don't have a social life some of these modules are really tough I think a couple of people take this part-time masters with me as well it's like 30-40 hours a week so my weekends are just at home my weekdays, every night if I manage to get home from my start-up job I was in the start-up, I'm now in the start-up if I can afford some time I'll squeeze in a lecture or two or write some code before it collapse but it's not easy if I had known this would be like this maybe at that point in time I would seriously have second thoughts it's quite difficult and your travel schedule is basically around my exams and let's travel you probably can't travel much but not to scare anyone off doing full-time masters it's very rewarding which masters you're taking is this college in US I don't know if you're up with Georgia Tech it's not bad, computer-sized college the profs are really good the education is really good as well okay, I'll touch on the next question that a lot of people ask okay is Ph.D. essential for your job role or not so I'll let we enough first so I didn't get to talk about the master because I already have a Ph.D so my Ph.D is the computational chemistry and I think it's quite necessary because for my current role because the basic knowledge has been covered mostly in my Ph.D. study that time I have learned the area and the curve and to learn what is great and decent all of those optimization methods but if you are if you want to take a Ph.D first of all you have to check for example there's IVM forecast to receive five years later what is the job market look like maybe by the time you graduate the job market has changed now I think the demand is much greater than the supply for data scientists and for recent years still should keep this I don't know five years later just keep yourself updated and not miss a good chance master is suggested I think to do the computer science and during the master you can still do some internship when you do some internship in a company they tend to want to break up you if you have some decent result so that's a good chance but Ph.D unless you really want to take up some role really scientific my role is already scientific enough if you want to take up my role and you come to the interview and if I will interview you if you have bachelor or master I think that's still okay if you have some basic numerical statistics some biologically accurate some Python coding college coding ability familiar command line interface those are something that are basic you have to know and I think you have to have some projects that demonstrate you can really do something to impact the company that's enough the degree is not so essential I think a very different view on this I am of the view that you don't even need a college degree to get a job recently we've heard the news Google, Apple, IBM Bank of America they don't even require a job because they have a college degree and I recall seeing a list of contributors to TensorFlow or the top minds in data science most of them don't even have college degrees the best people I've hired or physical engineering or chemistry or operations research so I definitely don't think you need a masters don't think you need a PhD even without a proper college technical degree you can actually learn everything yourself thanks to the internet so that's my thing very so I would say PhD definitely not a requirement as an engineer but I'm not really sure if you need a college degree I can't imagine a scenario where you don't need that as a programmer if you don't know operating systems so I think it's quite essential sure don't think you need a PhD to be a business analyst you have a PhD creating visual chances but then I think there are some people who do that so I guess back to previous questions so as you work with as a data analyst your main customer is always the coders whether it's operations or marketing or sales or sea level management the business degree helps to get you the language that they're talking about that helps the masters skills further put it into good use PhD is not really that required okay so we move on to another and that is actually not really another aspect it's one of the tools I think some of you are interested in what kind of tools that you are currently using now and let's say what is the go-to tool at least for the beginners who wants to work in the data science field what's the go-to tool that you recommend them to pick up first okay so Michael, where do you start first? pick up excel first though seriously I've seen some part-timers come in and say could you do control sea control and then we ask her to draw a chart she started drawing rectangles of different length so I don't think anyone in this room is at that level so that you go that far down excel get you started straight down already of course you need to start picking up sequel and all the other stuff let's step by step approach like this there isn't really a formula that's all this wena for me kaitan, as I already told you kaitan, library is a panas namhai so for other things you haven't know in closer because currently machine learning has some things to deep learning deep learning is part of machine learning it's just a more powerful when you the scale is really large so since the deep learning comes up so quickly you have to learn some deep learning tools like kaitoche it's just like this is like a via animes thing pencil flow and kaitoche so I don't really do just a kaitoche that you can call something and they're just like kaitoche it's something like you will fight each other to death if one guy uses pencil flow no no it's all just matrix multiplication and a chain rule it's all the same so I would say at the least kaitan and SQL SQL most data engineering jobs are moving towards completely I would agree with most of the panas SQL to extract the data because you don't want to be the helpless data scientist meeting for someone to give you the data in the CSV and python to manipulate the data like what we might mention pandas then probably to explore it I don't want to run some machine learning and so on it's easy to use api pencil flow, kaitoche lgbm there's a lot of stuff up there just take what you need you don't need to learn the whole ecosystem overnight and one thing because I work quite a few different companies as data panas or bi every time you change the different company you're using a different tool set and you have to learn that one so if someone say spot fire to a blue power be a kick set just learn one of them whichever one your job currently requires so that's the one you need but I guess that's also what I'm trying to learn now is what are the free options available so that's super set you can basically create charts upload it and you don't need to pay the same amount so that helps so tool sizes python, SQL excel for data analysis nothing more than excel okay so we'll take one question from here any question for my audience alright yes I think it's supposed to be your job eh I'm wondering about the value of this competition platform such as Kaggle so do you guys take participation in this kind of competition to help you improve your scale or okay so do you guys take part in Kaggle and if you do like what did you gain from it can you summarize it as that so yeah I didn't want to go first he just said it doesn't apply to him sorry am I bad so focus on Jun okay I think it's a simple problem because I'm a data scientist and sometimes people ask me whether if I don't have a PhD can I join your role and there is yes so you just have to learn all the basics I just mentioned and join some Kaggle competition because Kaggle competitions and projects those data sets are cleaned by those companies and they do have some problem they want users to solve and they provide you the data that's a very good chance and you can connect the thoughts that everything you have learned maybe you didn't learn everything but you just connect everything you have learned and apply to that project maybe after one or two projects you have a sense and you really like to seal this career or after one or two projects that you have you have something to demonstrate the interview work what you can do even without a PhD so that's quite a platform okay Michael yeah so Kaggle helps if you are like a fresh graduate and you didn't have a work experience and if in the interview I ask you do you have any experience with the data set because they already gave you a problem so a lot of times I think is how you walk through that problem how you look into it the other one for us as in visuals there's also the other challenge is and only if they challenge you on visualizing data that really helps you because they give you data sets and you're trying to go through it so it helps but I think I've got a great learning experience so first of all you attempt it and then you realize the leader has a faster career and then you realize so I think that's a very great experience okay so continue with that so how would you think Google how do you spell the word bus would you like Google for it first yeah I don't have my phone with me it's B-A-S-T so just so for B-A-S-T on Google you should be able to get it you don't get a dictionary okay anyway I think this is a great question I think a lot of the panelists have shared how Kaggle is amazing yes Kaggle you get very beautiful clean data sets you get to practice your machine learning skills your ninja hacking skills immediately and the one thing about Kaggle is that you get immediate feedback you get immediate feedback and you get to compare yourself against your peers or against people who are taking part in the competition that itself is a very useful feedback like hey you know how well do I do if after 3 or 4 competitions you're still in the bottom 20% maybe you probably won't enjoy this so Kaggle is one of that one thing interesting story I actually know of a person who got hired based on Kaggle so there was this someone who did categorization challenge on Kaggle so he did a sharing at a meet up just like this he did a sharing at a data science actually meet up so one of the heads of data science or one of the people at an e-commerce company happen to see the presentation and then they like invited him for lunch and you know we have a product categorization problem and we saw that and this person he took part in the competition he took the effort to share the competition as well so whenever you Kaggle do spend an hour or 2, write up your approach log in, share it, make it publicly available so people can know how you actually perform the challenge they want to know your top process and so they invited him in and they say we saw your slides, can you present to us so the guy presented his slides he presented his top process methodology and the next day they sent in an offer so you can actually get a job through Kaggle though I think this guy was probably pretty damn lucky but let me add the flip side of things who here actively practices data science in your day-to-day job okay, for how many of you every night as well for how many of you is okay for you is machine learning more than 50% of your job anyone here, machine learning is more than 50% but you have a great job we need to talk for how many here machine learning is more than 30% of your job that's cool so what you see is that machine learning is a significant portion is the part that drives a lot of the value but it's only about 20-30% a lot of it is cleaning the data Kaggle doesn't give you an opportunity to do that a lot of it is figuring out how to join the data understand the data, explaining your results that's why you need to write write and get feedback so to do that you can actually get practice by working on real problems sharing it at meetups get feedback that's how you prove and you also at the same time you'll be a portfolio so taking about the Kaggle competition it's useful for practice but it's not it's not it's not it's not it's not a competition it's useful for practice but to really build a portfolio you need to document it, share it either in a blog post or anything any sharing is better than no sharing after Kaggle you can use Kaggle as part of your portfolio to get the role so we've spent a bit of time on the technical skills that's quite a bit more about the soft skills what kind of tricks do you think are desirable for people working in the role that you are in what kind of tricks is desirable let's start with Michael then specifically tricks i guess the most important one is that you will have a lot of human time oh i'm going to specialize in data i'm just going to look at the data a lot that's totally reversed a lot of times you're talking to business stakeholders you're talking to people who are in operations and telling that something's wrong at the warehouse or go fix it something wrong is this wrong the first trait you need to know is that you are very comfortable presenting numbers to people and especially in ways that they can understand it if you've explained that in skill terms sure you might get a few hits but it's not going to get you the most of the hits so you need to be very good at communicating what's the other data how you got the analysis out so i think that's the most critical trait i can think of Jenny, what about for ML engineers say understanding the data how the data leads to the business of the company so what each field means is the impact of that field of the business reporting and the data science side i would agree with Jenny i would agree with Jenny i would say the most important trait for data scientists, someone who uses data is humility humility with the data so let's say in the data literature you say this primary key, this second key you just join them and you expect your model to run well well you didn't know there's a one to many mapping and the job breaks in production because it uses so much memory so being humble of the data means not assuming you know about the data unless you actually dig into it yourself this is very important for data science because you as we know in building good data science products the secret sauce is in the data preparation the feature engineering so being humble of the data i think another one is being curious of the data as well so if python, i think 3-4 tahun lalu now it's python nix, who knows what and spark didn't come into play 5-6 tahun lalu so you have to be curious you have to be disciplined to pick this up on your own as well as apply them on your job i think those are really important i agree with you Jenny because the intellectual curiosity is always the most required or personal character character for some candidate because even you know the machine learning models superficially but then you apply it you always want to take some insights you want to do it better than other users or other people or you just want to go that one extra mile than the others so that can make you the excellent model another thing that i think is you have to be very perspirits sometimes the project is so difficult you want to drop it in the middle but the business are being there so you just have to really keep working on that until you inspired by something and you get the great improvement and great impact and that's what they look for from data scientists so continue with that just to share your personal experience how did all of you decide which role in data to specialise in so what is uniquely interesting to your job role i think let's just answer the first one which is how did all of you decide which role in data to specialise in that's a good question beyond us i haven't really specialised yet my mission is just to use data to add value if that requires me to do ETL to do data acquisition i'll do it if that requires me to build the machine learning product so someone can use it i'll do it if that requires me to write internal tooling to make it easier for junior data scientists to be more efficient i'll do that as well whatever it takes i think that everyone should adopt this mindset whatever it takes for you to use the data to make impact in the most meaningful way don't be afraid to take ownership of it even though it's not your role hey that's a data engineering role i should wait for them to do ETL for me but you know what, if they're so busy write some spark on your learning know-harm or whatever so methods or processing techniques you use to process your data make it into a package that other people can contribute to and can use, that way you make the whole team be more efficient i haven't decided on the role so you're still moving or not for me there are two ways one is very selfish where i find that if you look at systems like Hadoop or Spark they are very complex systems so learning them takes a while it's very challenging giving issues on them is even more challenging so just work itself is very interesting even if you don't consider about business impact if you do get a role where you can work on a product which lots of people use i think the combination of two is a very big thing so i hope you are the point same as Eugene i got started when one of my earlier jobs we received a report from a market research and i think we all knew that it's a lot of money but we knew the value of it as well because the director just pointed at one number and yeah, that's the number we need and i just left that and that point that's where i realised there is value out of creating data making sure that companies utilize the data to do maximum making sure that it can impose value out of it making an impact out of it not just charge the data so that's one area but i agree with Eugene it's very hard to define it as a specialisation because sometimes i do a little bit of detail on my own i have to set up a server or sometimes i realise i'm doing classification by hand so could i do a little small data science project to do it so that actually did it so that really helps it's a good product so that's related to those different roles in big data or data science so do you know there is a data science pyramid like there are a few layers the first one is software engineer like data engineer put data into some formatical analysis then there's data analysis and then there's a machine learning scientist so for me i don't have a lot of roles that i can pick because i'm not from the computer science i can't be software engineer and i can't be data engineer so i just have to be data analyst or data scientist and i just i mean for this data science pyramid like every level you got you need less people so i think both of us just like not so many roles open to us but i think for those every upper layer that based on this lower level but i think it's also a good thing because all of those people related can become the other role for example if you're a software engineer and you already have the coding skill then you can begin at least the opportunity and all of those are just like ecosystem of the data space and you can just take for you great, let's really take another question may i know once you guys about most challenges you have met but the other side is you have met people maybe you have missed some trouble maybe you can create data or you build a model cannot show your matches like you want to achieve 100% and another question is about because as i know data science have a lot of overlap even that they are engineers sometimes we do overlap how do you communicate with them how much sense the data how do we gonna do so let me get the question first so what is the most challenging part of the work is that correct let's say if you are in a particular role are you able to move to another role is it how do you coordinate with data engineers also okay the coordination part i think you should you can check okay let me answer the second question first coordination with the data engineers of the platform engineers or anyone else i think i'll answer this with a program paradigm i mean anyone here work with you understand the concept of interfaces the api application protocol interface so i think how we structure our teams is we work with interfaces and each interface there is a contract for example i say okay data science we will build this api and these are the inputs it takes this is the SLA we can do one query every 40 milliseconds so we work with this interface so we define all these interfaces ahead of time data engineers are supposed to provide us this data in the raw data with this schema that's our input and then we produce a model we produce an output which is the interface the api contract and the team will call our model and use it so once we have defined this interface we can go on and develop whatever whatever we need to do as long as we meet this contract so that way we can parallelize our development so that's the second question we actually just work through interfaces just like how you work with programming languages you call sklab api you already know that it needs a X it needs a target etc but the main problem is sometimes the api design and interface are very important someone need to come up with the interface for each team so i mean sometimes we will meet some problems api design the interface is very important but in the first time you do this no one know this api design what does it like so yes i think if you are trying to push the edge a lot of things are not quite developed yet we don't quite know the best practices but we nonetheless do the best we can i think there are certain interface design whereby you can have a lot of flexibility in adding new keywords or adding new parameters so we try to adopt that as much as possible i think we are optimistic that whatever we build can last at least 18 months but frankly speaking i would assume that every 2-3 years we need to do a huge refactoring or to rebuild everything from scratch and i think that's what we see in the 3a now it's as well onto your first question what's the most challenging part of the world i would say that technical issues is not a problem put enough time and effort into it enough googling you will be able to solve it the difficult part is you build something and other people don't understand they don't want to use it how do you convince them there's no right answer to that you can do a simulation if you did this you can run an AD test and show them the conversion rate but they don't understand that they don't use it okay then you need to educate them hey you know this actually what we are doing is math and statistics but if they don't care anyway and they're like hey you know what we just increased conversion last week we put it into production oh okay then they accept it a lot of times the problem is getting people to understand and what you do and getting them to accept it and to use it that itself there's no formula to it that's what i find the most challenging okay i think you view all the time right what we will do is focus on the first question no no no i'm not again about you what's the most challenging part of your job i would say working with other teams it's not a negative thing it's like if you want to try a change how do you influence the other team to make it happen so to give an example it's a e-commerce website and you want a new field to get ready so getting the front-end developers to add it to the HTML and working with all the product managers and convincing them why this is important why you need this to be tried so things like this these are more challenging than being reactive like if something fails and you have people that's every day job so you that's not necessary for me sometimes you can ask questions that that's an ad-hoc request oh we need to do this and you just spend a lot of hours smashing your head on the table until something useful comes up you're just staring at all the charts and visualising it by every single cut possible that's the difficult part then you realise that after a few years you realise that you should not just take the request from someone and go home and do it or go back to the test and do it what you need to do is make sure you ask why do you need this after I give you this, what are you going to do with it but sometimes you still keep the pokok where you just look at the data and nothing good comes out of it but you still just coach it the challenging part of my job is just to deploy the model I don't have a lot of engineering background so also the machine learning is now working out a lot of deep learning models and those models really need to be backed up by machine learning engineers so we do have enough machine learning engineers sometimes my colleagues just work on the engineering part and we really want to so maybe we have to really catch up that part and we have to learn for ourselves to help deploy the model and that's a challenging part alright so any questions? You'll take just one more question okay we'll take okay we'll take two okay yes so I think one of the hardest things when you're self studying and trying to make a career change or move into a more technical role is how do you know when you've learned enough how do you know when it's time to say okay it's time to stop studying and go get a job and I guess an easy answer to that is apply it and once you get a job and you know but kind of a follow up question is there's data scientist but it seems like a high achievement but are there lower levels of data science are there junior data scientist what are the rungs where you can get into the profession I guess it's my question you want to specifically target the question for any of the panelists or I'll choose for you or whoever thinks they have an answer for it okay because in real time I think it's good to just maybe get about two of the panelists would you like to choose the two how about okay okay thanks I would say there's a very easy solution to your question how do you know you're good enough you find someone who is in a position you want to be and you talk to them and ask them hey I know this I know that am I good enough if not what am I missing what are the key skills that I need to be a data scientist what do I need to learn what are the biggest challenges you face I think by virtue of the fact that you are attending this meetup you already know that is one way to answer the question so when I was trying to be a good data scientist what I would do is I would meet I would interview not interview at companies but interview heads of data science or rock star data scientist anyone here know Xavier Conor Chief Data Scientist at Data Robot, Kaggle Master for many years running I would ask him hey no what does it take to be a good data scientist a rock star data scientist the peers around you you find it a lot of times it's not because they are a python ninja or SQL master a lot of times they can apply the technical skills and get it to create value for the company the model gets into production and gets measured so I would say that ask them what it takes to be good enough and then check against your your own skills inventory whether you have it or not sorry thank you for the question for me if you are taking some course for example master or PhD or even ungraded the best suggestion is for you to take some internship in some company before you graduate but if you are just looking for some mid career change that's also possible after you finish some projects for example you follow the two Kaggle competitions and you finish those projects then like you said the best idea is to show some of the people already in that role okay I have done two projects and this is how I did it and let me correct you okay so for my role people mostly do like this so that's the most efficient day for you to procure yourself the toolkit to suit for that company and also different or different company but they may be assigned a totally different job and maybe there are different requirements for a toolkit then if you learn Python maybe they require R if you learn C++ and that's a lot of effort then maybe they require Python so don't learn too much things in the beginning just have a look of that job description as person in the company or not in that company or you know someone else in the other company similar role you can just cannot learn to add a bit on to that I would say that you are probably good enough if you can build a product and deploy it on your own server and just think that on your phone, you know I build this, is it cool if someone says it's pretty cool that's a great portfolio right there at least you have done it end to end you found data, you put a model you deploy it on some chip cloud server, sport instance and someone can actually touch it someone can feel it and there's a basic UI I think that's good enough for me to act right will be very active in the community and look for mentors like Eugene Kaising over there these are some of the mentors you can approach okay, last question thank you so much for sharing I just want to I'm quite curious about this because my leader tell me that data science is different from engineering the most important reason is engineering product can always be successful as long as you pay enough effort and time and money and something like that but data science product can fail they can fail because of very different reasons maybe the assumption is wrong maybe your data is not good so my mother-in-law have some very impressive failure experience in the data science career so I'm asking for the two data scientists okay, very very much yeah I would work as a data scientist so I'll end actually it doesn't fail but as an engineer I've failed maybe hundreds of times okay there was one time when I pushed a single line of conflict change and it brought down a 500 known product cluster okay but at least there's something you can point or resume and tell people I failed I will put it but that's a good story to tell I mean, I don't know if you recall the guy that brought down EWS there was one time EWS just went down all that time internet went down but EWS they fire him they say hey, you know what can you prevent this from happening next time so I think that failure is not failure is actually a point for learning for me, I failed 10 times for one success I've had at least for some of my previous teammates in my previous role every time they see me with a grumpy face I've seen AP testing not going so well so like what you said data science is very ill-defined software engineering is a very mature craft we have very good design patterns to solve very standard problems or at least they've done a good job making it very standard for data science we have classification, regression and learning but even among those there's still a lot of different patterns that you will see, etc so you're right, data science is party research that's what makes it difficult but that's what I find makes it fun my most impressive failure I think there was once I ran an AP test and people found out and they were like hey, you know I ran an AP test that cost us hundreds of millions in revenue and I'm like okay, I'm going to be fired but I didn't I just came up with a plan, okay the AP test didn't do so well here is my hypothesis of why it didn't do so well here is how I'm going to investigate it here is when I'm going to tell you my investigation here is how I'm going to improve the model and here is when you see the next AP test so I think the more failure the better it actually helps you okay so in real time I think we will have just one more last question and the last question will be for those who wants to join your role what's the key piece of advice you'll give them to join to take up your role I'll let Eugene start first then you mean to to take up be my role I would say 2 things, practice and through your practice you get a portfolio I mean I don't know how many of you here renovate your own house or your family renovating your house would you hire a contractor that has no portfolio probably not so that's why you need a portfolio, it could be a blog it could be something on github it could be a carrier competition with a nice write up it could be some website a portfolio is important and if you can actually develop a portfolio it shows that you have practice so just get a portfolio and you will have taken care of the rest so as Eugene said I think github is like very very important github or a new purchase project so you can find so many products are there you can find the project that is most interesting to you and try to make small changes so if you have small changes then you have a lot of senior developers looking at your code so that way you can improve Mike I agree with Eugene you need to get a portfolio going it doesn't have to be work it doesn't have to be a carrier data set some of my friends said that I like magic the card carrying so he did his own visualisations on that really fantastic the other part is if your portfolio is not enough you can actually start talking to random companies and say that I can do this for you or what's your biggest problem and then you start talking to them that's one area that you can start off with that's a great organisation in Singapore called data kind SG it's an NGO that helps other NGOs make use of their data I used to volunteer that I don't have time now but if you want to get your hands dirty with real life messy data really messy among other experts who can provide guidance and training in a 2 day hackathon or data dive or whatever that is the place you go Thank you for both of you provide such a great time for us to practice data cleaning for my role I think the most important thing is you can have some decent degree to demonstrate or you take some of those online courses to demonstrate you have the basic knowledge plus that if you need to demonstrate some projects you've done and because we really want to see those people can do something can impact the businesses To summarise be able to demonstrate that you know how to apply whatever that you have learnt okay I've been said Thank you very much for the panelists to be here Thank you Eugene Now I have to fulfil my last application and that is any of you are looking to hire your organisation is looking to hire Yes If you are really interested in health tech using data to improve health talk to me Okay Data engineers Okay Okay We are also looking out for people who can work well so we are looking for hybrids so it could be like I only learnt data mix between data engineering but you are looking out for clients Okay So are you single for also want to hire some data scientist Actually just knowing that I've come to analyst our HR team so follow me here today so if you want to talk with them later after this Maybe the HR team want to Ya Just wait So thank you very much for spending your time here I hope the panel discussion you gain a lot more information from that So with that being said Thank you very much for coming down here Also a big thank you to our sponsor for the venue Yes So the panel needs to still be around So if you want to ask them questions, please stay back and talk to them Just still run like that Okay, so thank you very much Thank you