Become a data scientist without a PhD

The business value of AI is projected to reach $1.2 trillion this year according to Gartner. Yet 40% of enterprise companies are not adopting AI. Why not?

3 simple reasons:

      1. Data Scientists are hard to find and hire. Want to become one? There’s a lab for that. Enroll in the Data Science on GCP quest by Friday, October 12th and you’ll get 40 credits (free of charge, no CC required). The labs in this quest are derived from the book Data Science on Google Cloud Platform.
      2. Data analysts with SQL skills (and other programming languages) are also rare. Get practice running SQL queries in the BigQuery console in the Introduction to SQL for BigQuery and Cloud SQL lab. Then experiment with SQL and ML in the Ingest data into the Cloud Using Google App Engine lab. You’ll run Python scripts to download, automatically fetch, and clean data using Google App Engine. You’ll also create a new application and deploy it to the Google App Engine:

Then use a Flask framework to ingest data and invoke it using cron:3. Infrastructure AI is resource-intensive, in terms of both staff, and compute power. Many on-prem setups simply cannot handle the demands of AI. You can learn to take advantage of Google Cloud’s compute power to run your advanced AI jobs. The Google Cloud Dataflow to process data lab shows you how to configure BigQuery and install Python packages to use Apache Beam:

Then you will monitor the progress of your Cloud Dataflow job and inspect the processed data:

You’ll also:

Need more practice with the GCP infrastructure?  Visualize data with Google Data Studio by running the query to get the IP address to connect to the Cloud SQL:

Then create table views to look at flights that are delayed by 10, 15, 20 minutes:

After connecting with the Data Studio, you’ll create a data visualization for flight delays:

Tackle the quest to practice ingestion, preparation, processing, querying, exploring and visualizing of data sets using GCP. When you earn the badge, let us know. You can find us at @Qwiklabs – and we’re always happy to hear about your accomplishments!

Dig deep into BigQuery Machine Learning with Google Cloud

You don’t need a B.S. in Computer Science to take advantage of powerful GCP data analysis tools. Google announced fully managed service at Next’18, BigQuery Machine Learning (BQML). BQML allows you to:

  • Create, train, evaluate, predict and deploy machine learning models with minimal coding
  • Use SQL directly in the database

Get started with BQML by taking the Predict Visitor Purchases with a Classification Model in BQML lab today and get a 30 day pass (free of charge, no CC required) to help you take advantage of this shiny new toy — I mean, seriously, an impactful data analysis tool!

Here’s some of what you’ll do in the lab:

  • Explore Ecommerce data & run the Query to find out what percentage made a purchase from website visitors:
  • Run the Query to find out your top 5 selling products: 
  • Preview the demo dataset to find useful features a machine learning model uses to understand if a first time visitor will return to make a purchase & dig deeper to know the risks of only using these fields for your classification model:
  • Evaluate classification model to maximize the True Positive Rate (predict that returning users will make a purchase) & visualize a ROC (Receiver Operating Characteristic) curve to maximize the area under the curve:Take the lab  and get your 30 day pass to be one step closer to earning the Data Engineering Quest badge.