Tweet
Tweet

Slide 1

Slide 1 text

No content

Slide 2

Slide 2 text

No content

Slide 3

Slide 3 text

Code Along session - members Kamolphan Liwprasert (Fon) Cloud GDE and WTM Ambassador Burasakorn Sabyeying (Mils) Cloud GDE and WTM Ambassador Kasin Sumalyaporn (Up) Customer Solutions Consultant @ Google Punsiri Boonyakiat (Beat) Senior Data Engineer @ CJ Express

Slide 4

Slide 4 text

No content

Slide 5

Slide 5 text

No content

Slide 6

Slide 6 text

No content

Slide 7

Slide 7 text

Google I/O Extended Bangkok - June 22 Google I/O Cloud Extended Bangkok - Soon! Facebook Page: GDG Bangkok Facebook Group: GDG Cloud Bangkok

Slide 8

Slide 8 text

facebook.com/groups/gdgcloudbkk gdg.community.dev/ gdg-cloud-bangkok/

Slide 9

Slide 9 text

Burasakorn Sabyeying (Mils) Senior Data Engineer @ CJ Express, Google Developer Experts in Cloud goo.gle/ChaiyoGCPS4

Slide 10

Slide 10 text

Educating Developers Through Study Jams

Slide 11

Slide 11 text

Self-paced learning. Learn in your own time and schedule Gamification. Get the knowledge and score the swags! From community to community. Learning together with the community ChaiyoGCP: Online upskilling campaign for the developer communities to connect and learn about Google's latest technology. 1 2 3 goo.gle/ChaiyoGCPS4

Slide 12

Slide 12 text

March Follow up Wrap up Code Along session Start Event Recording in Youtube: ‘GDG Cloud Bangkok’ channel Social Sharing Hashtag #ChaiyoGCP Welcome Email: “[#ChaiyoGCP] ยินดี ต้อนรับสู่ #ChaiyoGCP! เริ่ม ต้นการผจญภัยแห่งการเรียนรู้ ของคุณได้แล้วตอนนี้!” Redeem 1 month access on Cloud Skill Boosts Quest: Generative AI Explorer - Vertex AI Code Along sessions (Online event) Completion Form Swags (6-7 weeks after result email) Make your profile public Google I/O Extended ( June 22 ) Result email will be sent within 1 week May 17 March April May 28 June 15 Stay Tuned!! ! !! Quest: Gemini for Data Scientists and Analysts Join Discord Community *Only available for developers and participants who live in Thailand.

Slide 13

Slide 13 text

No content

Slide 14

Slide 14 text

Build with AI + ChaiyoGCP (Kick off) May 19, 2024

Slide 15

Slide 15 text

Google Cloud Skills Boost

Slide 16

Slide 16 text

Tracks Quest/Badge (get badge here!) Lab vs Quest vs Badge Lab AI/ML Infrastructure & Security goo.gle/ChaiyoGCPS4 Quest/Badge Lab Lab Lab Lab Lab Lab Quest/Badge Lab Lab Lab Lab Quest/Badge Lab Lab Lab Lab Quest/Badge Lab Lab Lab Lab Lab Quest/Badge … Quest/Badge …

Slide 17

Slide 17 text

No content

Slide 18

Slide 18 text

List of Quests goo.gle/ChaiyoGCPS4

Slide 19

Slide 19 text

List of Quests Example: Tier 1 goo.gle/ChaiyoGCPS4

Slide 20

Slide 20 text

List of Quests Example: Tier 2 goo.gle/ChaiyoGCPS4

Slide 21

Slide 21 text

No content

Slide 22

Slide 22 text

FAQ Q: Where is the Referral Code and How is the Referral Criteria? A: You will get it in the welcome email. You can use that code to invite your friends to the ChaiyoGCP program and ask them to submit the Completion form. Success Referral Criteria: ● Your friends should put your referral code on register form or on their completion form (case sensitive) ● Your friends should complete at least 7 quests ● We will be selecting the top 3 ChaiyoGCP participants with the most successful referral to win a secret special prize . ● The winner will be contacted through email two weeks after the program ends.

Slide 23

Slide 23 text

FAQ Q: Can I mix the quests between two tracks (AI/ML and Infra+security)? A: Yes, you can! Q: If I have done the quest before, how to count complete missions? A: Only quests completed by 17 May - 15 June 2024 will be considered in the redemption process.

Slide 24

Slide 24 text

FAQ Q: Can I use the same email I used in last year ? A: Yes, you can. But the quest that already completed in last year won’t be count in this year.

Slide 25

Slide 25 text

Reason participant not getting swag ● You are not based in Thailand ● You completed fewer than the minimum of 7 required quests (inc. at least 3 skill badges) (List of quests) ● You did not submit your Google Cloud Skills Boost Profile public profile link ● You submitted an invalid Google Cloud Skills Boost Profile public profile link ● Your completion form submission passed the deadline

Slide 26

Slide 26 text

Tips & Tricks Use incognito window for lab login Incognito window Normal window

Slide 27

Slide 27 text

Kickoff Event (Build with AI day) How to redeem Cloud Skill boost - 1 month access & Get 9 credits 8 credits Monthly subscription Any 1 lab in this quest ** Open link in incognito mode** ** and ensure to sign out Cloud Skill Boost before click link** Join & Sign in

Slide 28

Slide 28 text

Today’s Lab !

Slide 29

Slide 29 text

Have fun !

Slide 30

Slide 30 text

Kasin Sumalyaporn (Up) Customer Solutions Consultant @ Google [LAB] Analyze data with Gemini assistance

Slide 31

Slide 31 text

Distributed Memory Shuffle Tier BigQuery Replicated, Distributed Storage (99.9999999999%) High-Available Cluster Compute (Dremel) BigQuery’s serverless architecture enables sophisticated compute capabilities Decoupled storage and compute Fully managed and serverless for maximum agility and scale Gigabyte to petabyte scale storage and SQL queries Built-in ML for out-of-the-box predictive insights

Slide 32

Slide 32 text

Gemini in BigQuery Always-on Intelligence spanning the Data & Analytics journey Explore, Discover & Analyze Engineer & Transform Optimize Workloads Secure Data Fundamentals

Slide 33

Slide 33 text

Gemini in BigQuery Engineer & Transform Data preparation Preview Explore, Discover & Analyze Data Canvas Preview Semantic metadata search Preview Data insights Preview Fundamentals SQL code assist Preview Python code assist Preview Accelerated migrations Preview Secure Data Assisted data masking Roadmap Privilege anomaly detection Roadmap Sensitive data monitoring Roadmap Optimize Workloads Partitioning & clustering recommendations Roadmap Materialized views recommendations Roadmap Autotuning for Spark Preview Assisted Spark troubleshooting Preview

Slide 34

Slide 34 text

This Lab: Analyze data with Gemini assistance

Slide 35

Slide 35 text

Issues: Streaming/batch data Data processing Export data Manage resources? Data Governance? Deploy ML model Train ML model (e.g. Python/R) Where do I host? Multiple products & roles can lead to unnecessary complexity & costs BigQuery (data warehouse) Typical ML workflow poses many challenges

Slide 36

Slide 36 text

Issues: Streaming/batch data Data processing Export data Manage resources? Data Governance? Deploy ML model Train ML model (e.g. Python/R) Where do I host? Simplify with BigQuery ML BigQuery (data warehouse) BigQuery ML greatly simplifies the ML workflow

Slide 37

Slide 37 text

CREATE OR REPLACE MODEL `bqml.penguins_model` OPTIONS (model_type='linear_reg', input_label_cols=['body_mass_g']) AS SELECT * FROM `public-data.ml_datasets.penguins` WHERE body_mass_g IS NOT NULL Train and deploy ML models in SQL Execute ML workflows without moving data from BigQuery Built-in infrastructure management, security & compliance ML with just a few SQL statements SELECT * FROM ML.PREDICT ( MODEL`bqml_tutorial.penguins_model`, (SELECT * FROM`public-data.ml_datasets.penguins` WHERE body_mass_g IS NOT NULL AND island = "Biscoe")) Train Predict 1 2

Slide 38

Slide 38 text

Enterprise benefits of BigQuery ML Cost & Complexity BigQuery’s fully managed and serverless infrastructure significantly reduces cost and complexity Security IAM roles and fine-grained access controls for all users on one platform People Business/Data Analysts can build and deploy powerful ML models using SQL

Slide 39

Slide 39 text

Lab Objectives ใน Lab วันนี้เราจะไดลงมือตาม Task ดังตอไปนี้ ● ใช Gemini ถามตอบคําถาม เพื่อทําความเขาใจโซลูชันและบริการตางๆ บน Google Cloud และการนําไปใช ● ใช Gemini อธิบาย SQL query ที่มีอยูแลว และเขียน SQL query ขึ้นใหมโดยใช Gemini เปนตัวชวย ● สรางและเทรนโมเดล Machine Learning โดยใช ARIMA PLUS (time series forecast) เพื่อทํานายยอด ขายในอนาคต ผานการเขียน SQL บน BigQuery (BigQuery ML)

Slide 40

Slide 40 text

Task 1. Configure your environment and account ● Enable the Cloud AI Companion API for Gemini ● Grant the necessary IAM roles to your Student Account and Quicklab Service Account Reference: https://console.cloud.google.com/apis/library/cloudaicompanion.googleapis.com gcloud services enable cloudaicompanion.googleapis.com --project ${PROJECT_ID} gcloud projects add-iam-policy-binding ${PROJECT_ID} --member user:${USER} --role=roles/cloudaicompanion.user gcloud projects add-iam-policy-binding ${PROJECT_ID} --member user:${USER} --role=roles/serviceusage.serviceUsageViewer

Slide 41

Slide 41 text

No content

Slide 42

Slide 42 text

Task 2. Task 2. Create a dataset and enable Gemini features in BigQuery (1/2)

Slide 43

Slide 43 text

Task 2. Task 2. Create a dataset and enable Gemini features in BigQuery (2/2)

Slide 44

Slide 44 text

Task 3. Use Gemini to analyze your data

Slide 45

Slide 45 text

Task 4. Prompt Gemini to explain SQL queries in a sales dataset

Slide 46

Slide 46 text

Task 5. Generate a SQL query that groups sales by day and product # select the sum of sale_price by Date(created_at) and product_id casted to day from bigquery-public-data.thelook_ecommerce.order_id as t1 joined this with products table in the same dataset as t2 SELECT SUM(sale_price), DATE(created_at) AS created_at_day, CAST(product_id as INT64) FROM `bigquery-public-data.thelook_ecommerce.order_items` AS t1 JOIN `bigquery-public-data.thelook_ecommerce.products` AS t2 ON t1.product_id = t2.id GROUP BY created_at_day, product_id

Slide 47

Slide 47 text

Task 6. Build a forecasting model and view results CREATE MODEL bqml_tutorial.sales_forecasting_model OPTIONS(MODEL_TYPE='ARIMA_PLUS', time_series_timestamp_col='date_col', time_series_data_col='total_sales', time_series_id_col='product_id') AS SELECT sum(sale_price) as total_sales, DATE(created_at) as date_col, product_id FROM `bigquery-public-data.thelook_ecommerce.order_items` AS t1 INNER JOIN `bigquery-public-data.thelook_ecommerce.products` AS t2 ON t1.product_id = t2.id GROUP BY 2, 3;

Slide 48

Slide 48 text

Bring AI to your data with the simplicity and scale of BigQuery BigQuery ML: Democratize machine learning Data to AI Inference Engine for remotely hosted models and Google’s pretrained models. Register your BigQuery ML models in the Vertex AI Model Registry Modeling capabilities ARIMA+, Explainable AI, Advanced Feature Engineering, Holiday modeling Model Ops Use Colab notebooks to perform ML workflows in BigQuery. Import tensorflow models for batch prediction or export models from BigQuery ML for online prediction Predict sales figures Predict stock prices Identify spam emails Identify tumor types Product recommendation Create personalized content Analysis of written text Analysis of DNA data Detect fraud Predict credit risk Perform customer segmentation Predict housing prices based on historical data Autoencoder K-Means ARIMA-PLUS PCA Matrix Factorization Wide & deep classifier DNN Classifier AutoML table classifier Boosted trees classifier Logistic regression Wide & deep regressor DNN regressor AutoML Table regressor Boosted Trees regressor Linear regression Predict values Predict between categories Generate recommendations Reduce data dimensionality Find anomalies Group data into clusters Time series forecasting What do you want to do? Legend Start Tasks Models Examples

Slide 49

Slide 49 text

Visual & AI driven data discovery, modeling, and analysis Gemini in BigQuery BigQuery Data Canvas GenAI-centric experience for data exploration and visualization Iterative and guided user experience Embedded inside BigQuery Studio Semantic data discovery supported by Dataplex catalog Automated python notebook generation Built-in collaboration for data analysts

Slide 50

Slide 50 text

Google Cloud Proprietary & Confidential

Slide 51

Slide 51 text

Punsiri Boonyakiat (Beat) Senior Data Engineer @ CJ Express [LAB] Gemini for Data Scientists (+Quiz)

Slide 52

Slide 52 text

Recap: Gemini for Data Analysis

Slide 53

Slide 53 text

This Lab: Gemini for Data Scientists Colab Enterprise

Slide 54

Slide 54 text

Lab Objectives - Gemini for Data Scientists ใน Lab วันนี้เราจะไดลงมือตาม Task ดังตอไปนี้ ● ใชงาน Colab Enterprise Python notebooks ที่อยูใน BigQuery Studio. ● ใชงาน BigQuery DataFrames inside of BigQuery Studio. ● ใชงาน Gemini Code Assistance เพื่อชวยในการ generate code ในการสราง Model และวิเคราะห ขอมูล ● สราง K-means clustering model เพื่อชวยในการจัดกลุมของขอมูล ● ใช Vertex AI - text-bison model เพื่อชวยวิเคราะหกลุมของขอมูลลูกคา E-commerce เพื่อใชงานการ สราง marketing campaign

Slide 55

Slide 55 text

Task 1. Configure your environment and account ● Enable the Cloud AI Companion API for Gemini ● Grant the necessary IAM roles to your Student Account and Quicklab Service Account Reference: https://console.cloud.google.com/apis/library/cloudaicompanion.googleapis.com

Slide 56

Slide 56 text

Task 1. Configure your environment and account c ● Enable the Cloud AI Companion API for Gemini ● Grant the necessary IAM roles to your Student Account and Quicklab Service Account

Slide 57

Slide 57 text

Task 2. Enable New code management features c

Slide 58

Slide 58 text

Task 3. Create a BigQuery dataset for your project

Slide 59

Slide 59 text

Task 4. Create a new Python notebook and connect to the runtime

Slide 60

Slide 60 text

Task 5. Build the Python Notebook ● Import Python libraries ● Define variables ● Create and import a base table as a BigQuery DataFrame from a public dataset ● Generate the K-means clustering model and visualization

Slide 61

Slide 61 text

Understand Data - theLook eCommerce theLook eCommerce dataset documentation: https://cloud.google.com/bigquery/public-data/thelook

Slide 62

Slide 62 text

What is BigFrames ? References: ● https://github.com/googleapis/python-bigquery-dataframes ● https://cloud.google.com/python/docs/reference/bigframes/latest

Slide 63

Slide 63 text

Python SDK for Gemini API in Vertex AI

Slide 64

Slide 64 text

No content

Slide 65

Slide 65 text

Understand Data - theLook eCommerce

Slide 66

Slide 66 text

Create Customer statistic with BigQuqery magic command Reference: https://cloud.google.com/python/docs/reference/bigquery/latest/magics

Slide 67

Slide 67 text

K-Means Clustering https://www.naftaliharris.com/blog/visualizing-k-means-clustering/

Slide 68

Slide 68 text

Check Customer Segmentation Model ! Model Predictions Results

Slide 69

Slide 69 text

Cluster Summary Task 6. Generate insights from the results of the model

Slide 70

Slide 70 text

Task 6. Generate insights from the results of the model Example: Cluster 1: Title: "The Occasional Shoppers" Persona: These customers are likely to be sporadic shoppers who make infrequent purchases. They may be attracted to discounts or promotions, and they may be more likely to purchase items that are on sale. Next marketing step: Offer discounts or promotions to entice these customers to make more frequent purchases.

Slide 71

Slide 71 text

Quiz Time!

Slide 72

Slide 72 text

Congratulation! ใน Lab วันนี้เราไดลงมือทําไปดวยกัน!! ✌ ● ✅ ใชงาน Colab Enterprise Python notebooks ที่อยูใน BigQuery Studio. ● ✅ ใชงาน BigQuery DataFrames inside of BigQuery Studio. ● ✅ ใชงาน Gemini Code Assistance เพื่อชวยในการ generate code ในการ สราง Model และวิเคราะหขอมูล ● ✅ สราง K-means model จาก BigFrames ML เพื่อชวยในการจัดกลุมของ ขอมูล ● ✅ ใช Vertex AI - text-bison model เพื่อชวยวิเคราะหกลุมของขอมูลลูกคา E-commerce เพื่อใชงานการสราง marketing campaign

Slide 73

Slide 73 text

แนะนํา Gemini Learning Path ✅ https://www.cloudskillsboost.google/paths/236