Earlier this year, we launched the Google Cloud skills challenge, which provides 30 days of free access to training to build your cloud knowledge and an opportunity to earn skill badges that showcase your Google Cloud competencies. Today, we’re adding a Native App Development track to the skills challenge, joining the Getting Started, Data Analytics, Kubernetes, Machine Learning (ML) and Artificial Intelligence (AI) tracks. 

The Native App Development track is designed for cloud developers who want to learn to build serverless web apps and Google Assistant applications on Google Cloud using Cloud Run and Firebase. Specifically, you’ll have an opportunity to earn three skill badges in the Native App Dev track: Serverless Firebase Development, Serverless Cloud Run Development, and Build Interactive Apps with Google Assistant. To earn a skill badge, you complete a series of hands-on labs and take a final assessment challenge lab to test your skills.

Here’s an overview of each badge.

Serverless Firebase Development

To earn this skill badge, you’ll learn how to build serverless web apps, import data into a serverless database, and build Google Assistant applications using Firebase, Google’s backend-as-service platform for creating mobile and web applications.

Serverless Cloud Run Development

For this badge, you’ll discover how to use Cloud Run, a fully managed serverless platform, to connect and leverage data stored in Cloud Storage. You’ll learn how to use Cloud Run to build a resilient, asynchronous system with Pub/Sub, build a REST API gateway as well as build and expose services. 

Build Interactive Apps with Google Assistant

To earn the final skills badge, you’ll build Google Assistant applications by creating a project in the Actions console, integrating Dialogflow, testing your action in the Actions simulator, and adding Cloud Translation API to your assistant application. 

Ready to jump into the skills challenge? Sign up here. 

You can also check out this quick video below to learn how to join the skills challenge.

Cloud Spanner is a fully managed relational database with unlimited scale, strong consistency, and up to 99.999% availability. It is designed for highly concurrent applications that read and update data, for example, to process payments or for online game play. To ensure the consistency across multiple concurrent transactions, Cloud Spanner uses a combination of shared locks and exclusive locks to control access to the data. In this blog, we will explore the different types of locks present in Cloud Spanner. We will also discuss some common cases of transactional locking in Cloud Spanner, and what to look out for to detect when these cases might be occurring.

What is a lock?

Before we get into the details, let us first quickly recap and define a lock in the context of database systems.

“Locks” in databases are a mechanism for concurrency control. Locks are typically held on a resource, which may mean rows, columns, tables or even entire databases. When a resource is locked by a transaction, it cannot be accessed by another transaction until the lock is released.

Timeline view of transactions

Before we proceed to discuss transaction locking in the context of read and read-write transactions, it is important to recap the timeline view of the transactions in Spanner.

Please also keep in mind the concept of write buffering in transactions, since we refer to it in the Common Transaction Patterns section below. Write buffering refers to the Cloud Spanner server(s) accepting the writes. Note that these writes are not durable until a commit has been performed.

Types of locks in Cloud Spanner

Cloud Spanner operations acquire locks when the operations are part of a read-write transaction. Read-only transactions do not acquire locks. Unlike other approaches that lock entire tables or rows, the granularity of transactional locks in Spanner is a cell, or the intersection of a row and a column. This means that two transactions can read and modify different columns of the same row at the same time. To maximize the number of transactions that have access to a particular data cell at a given time, Cloud Spanner uses different lock modes.

Here is a brief description of the different lock types. Learn more about each lock type in the Cloud Spanner documentation.

1. ReaderShared Lock – Acquired when a read-write transaction reads data.
2. WriterShared Lock – Acquired when a read-write transaction writes data without reading it.
3. Exclusive Lock – Acquired when a read-write transaction which has already acquired a ReaderShared lock tries to write data after the completion of read. It is a special case for a transaction to hold both the ReaderShared lock and WriterShared lock at the same time.
4. WriterSharedTimestamp Lock – Special type of lock acquired when inserting new rows with the transaction’s commit timestamp as part of the primary key.

Handling Lock Conflicts

Since read-write transactions use locks to execute atomically, they run the risk of deadlocking. For example, consider the following scenario: transaction Txn1 holds a lock on record A and is waiting for a lock on record B, and Txn2 holds a lock on record B and is waiting for a lock on record A. The only way to make progress in this situation is to abort one of the transactions so it releases its lock, allowing the other transaction to proceed.

Cloud Spanner uses the standard “wound-wait” algorithm to handle deadlock detection. Under the hood, Spanner keeps track of the age of each transaction that requests conflicting locks. It also allows older transactions to abort younger transactions, where “older” means that the transaction’s earliest read, query, or commit happened sooner.

Common Transaction Patterns

Armed with the basics of transactions and locks in Cloud Spanner, we will now walk through a few practical use-cases. We take the example of an application which queries and updates users’ balance in a table named accounts. The accounts table has the following columns –

Case 1: Transaction waiting to get exclusive lock because of higher priority shared-lock

Sequence

1. Txn1 begins.
2. Txn2 begins.
3. Txn2 reads the table and acquires a ReadShared Lock.
4. Txn1 buffers its write.
5. Txn1 tries to commit, but since Txn2 has higher priority (because it has executed its first operation first), Txn1 will have to wait until Txn2 releases the ReadShared Lock.

After committing, Txn2 releases its ReadShared Lock. Txn1 is now able to upgrade to an Exclusive Lock. It acquires the Exclusive Lock and commits.

Note 1: UPDATE WHERE is always transformed into SELECT WHERE, so an UPDATE statement is not a blind write, but a read-write operation.

Note 2: While the above example considers the use-case of querying and updating a single row, the same transaction pattern is also applicable across a key-range.

What to watch out for

• Timeouts due to Transactions waiting for acquiring locks. Refer to transaction statistics and lock statistics to understand how to detect this.

When does this typically happen

• Concurrent updates to a single key/key-range in a table.

Case 2: Transaction aborted because of concurrent execution succeeding

Sequence

1. Txn1 begins.
2. Txn2 begins.
3. Txn1 reads a row from the table and acquires a ReadShared Lock.
4. Txn2 reads the same row as Txn1 and acquires a ReadShared Lock.
5. Txn1 buffers its write.
6. Txn2 buffers its write.
7. Txn1 tries to commit, has higher priority (because it has executed its first operation first), it will get priority in upgrading to an Exclusive Lock. It acquires the Exclusive Lock and commits.

Since Txn1 has committed and was the higher priority transaction, Txn1 will abort Txn2.. The abort will likely happen before Txn2 tries to commit.

What to watch out for

• High number of transaction aborts due to wounding of transactions. Refer to transaction statistics and lock statistics to understand how to detect this.

When does this typically happen

• Concurrent updates to a single key/key-range. Typically this happens in the case of hot keys being present in the table.

Case 3: Transaction waiting to get exclusive lock because of higher priority shared-lock & getting aborted because of prior succeeding concurrent execution

Sequence

1. Txn1 begins.
2. Txn2 begins.
3. Txn2 reads the table and acquires a ReadShared Lock.
4. Txn1 reads the table and acquires a ReadShared Lock.
5. Txn1 buffers its write.
6. Txn2 buffers its write.
7. Txn1 tries to commit, but since Txn2 holds a ReadShared Lock, it has to wait until it gets cleared.
8. Since Txn2 has higher priority (because it has executed its first operation first), it will acquire the Exclusive Lock first. Txn2 upgrades its ReaderShared Lock to an Exclusive Lock and commits.

Finally, after Step 8, when Txn1 tries to commit (since ReaderShared Lock from Txn2 is now cleared) it gets aborted. This is done to prevent deadlock with the higher priority transaction (Txn2).

Note: In Case 1, even though Txn2 had acquired a ReaderShared Lock earlier as in Case 3, it never upgraded to an Exclusive Lock (since there were no writes). Hence there was no need to abort Txn1, it could simply wait for Txn2 to release its ReaderShared lock and then commit.

What to watch out for

• High number of transaction aborts due to high contention between transactions. Refer to transaction statistics and lock statistics to understand how to detect this.

When does this typically happen

• Concurrent updates to a single key/key-range. Typically this happens in the case of hot keys being present in the table.

Recommendations

In order to mitigate these issues and reduce their occurrence. We recommend adopting the following best practices:

• If you need to perform more than one read at the same timestamp, and know in advance that you only need to read, consider using a read-only transaction. Because read-only transactions don’t write, they don’t hold locks and they don’t block other transactions. Further, read-only transactions never abort, so you don’t need to wrap them in retry loops.
• Always acquire ReadShared Locks on the smallest subset of keys or key ranges. This reduces the chances of lock contention.
• Analyze your code to only include critical path code within a transaction. Avoiding unneeded remote calls or complex, long-running business logic is a good way to ensure a transaction process quickly.
• Analyze your needs for multi-split transactions. Since transactions that update more than one split use a 2-phase commit protocol, they hold locks for a longer duration, thereby increasing chances of lock contention.

Get started today

Spanner’s unique architecture allows it to scale horizontally without compromising on the consistency guarantees that developers rely on in modern relational databases. Try out Spanner today for free for 90 days or for as low as $65 USD per month.

Over the past two decades, consumers have become accustomed to receiving personalized recommendations in all facets of their online life. Whether that be recommended products while shopping on Amazon, a curated list of apps in the Google Play store, or relevant videos to watch next on YouTube. In fact, in a Verge article “​​How YouTube perfected the feed: Google Brain gave YouTube new life,” the Google Brain team reveals how their recommendation engine has impacted the platform with “more than 70 percent of the time people spend watching videos on the site being driven by YouTube’s algorithmic recommendations” thereby increasing time spent on the platform by 20X in three years. 

It’s become clear that personalized recommendations are no longer a differentiator for an organization but rather something consumers have come to expect in their day-to-day experiences online. So what should you do if you are behind the curve and want to get started or simply want to improve upon what you already have? While there are all sorts of techniques, from content-based systems to deep learning methods, our goal in this recommender-focused blog series is to demystify three available approaches to building recommendation systems on Google Cloud: Matrix Factorization in BigQuery Machine Learning (BQML),  Recommendations AI, and deep retrieval techniques available via the Two-Tower built-in algorithm.

One of these approaches can be used to meet you where you are in your personalization journey, no matter if you are just starting or if you are well into it. This first blog post will introduce our three approaches and when to use them. 

What is Matrix Factorization and how does it work?

Collaborative filtering is a foundational model for building a recommendation system as the input dataset is simple and the embeddings are learned for you. How does Matrix factorization fit into the mix you might be wondering?  Matrix factorization is simply the model that applies collaborative filtering. BQML enables users to create and execute a matrix factorization model by using standard SQL directly in the data warehouse. 

Collaborative filtering begins by creating an interaction matrix. The interaction matrix represents users as a row and items as columns in your dataset. This interaction matrix often is sparse in nature as not all users will have interacted with many items in your catalog. This is where embeddings come into play. Generating embeddings for users and items not only allows you to collapse many sparse features into a lower dimensional space but they also allow you to derive a similarity measure so that similar users/items fall nearby in the embedding space. These similarity measures are key as collaborative filtering uses similarities between users and items to make the end recommendations. The underlying assumption being that similar users will like similar items whether that be movies or handbags. 

What’s required to get started?

To train a matrix factorization model you need a table that includes three input columns: user(s), item(s), and an implicit or explicit feedback variable (e.g., ratings is an example of explicit feedback). With the base input dataset in place, you can then easily run your model in BigQuery after specifying several hyperparameters in your CREATE MODEL SQL statement. Hyperparameters are available to specify the number of embeddings, the feedback type, the amount of L2 regularization applied and so on.    

Why use this approach and who is it a good fit for?  

As mentioned earlier, Matrix Factorization in BQML is a great way for those new to recommendation systems to get started. Matrix factorization has many benefits: 

• Little ML Expertise: Leveraging SQL to build the model lowers the level of ML expertise needed
• Few Input Features: Data inputs are straightforward, requiring a simple interaction matrix
• Additional Insight: Collaborative filtering is adept at discovering new interests or products for users 

While Matrix Factorization is a great tool for deriving recommendations it does come with additional considerations and potential drawbacks depending upon the use case. 

• Not Amenable to Large Feature Sets: The input table can only contain two feature columns (e.g., user(s), item(s)). If there is a need to include additional features such as contextual signals, Matrix factorization may not be the right method for you.  
• New Items: If an item is not available in the training data, the system can’t create an embedding for it and will have difficulty recommending similar items. While there are some workarounds available to address this cold-start issue, if your item catalog often includes new items, Matrix factorization may not be a good fit.  
• Input Data Limitations: While the input matrix is expected to be sparse, training examples without feedback can cause problems. Filtering for items and users that have at least a handful of feedback (e.g., ratings) examples can improve the model. More information on limitations can be found here. 

In summary, for users with a simplified dataset looking to iterate quickly and develop a baseline recommendation system, Matrix Factorization is a great approach to begin your personalization AI journey. 

What is Recommendations AI and how does it work?

Recommendations AI is a fully managed service which helps organizations deploy scalable recommendation systems that use state-of-the-art deep learning techniques, including cutting-edge architectures such as two-tower encoders, to serve personalized and contextually relevant recommendations throughout the customer journey.

Deep learning models are able to improve the context and relevance of recommendations in part because they can easily address the previously mentioned limitations of Matrix Factorization. They incorporate a wide set of user and item features, and by definition they emphasize learning successive layers of increasingly meaningful representations from these features. This flexibility and expressivity allows them to capture complex relationships like short-lived fashion trends and niche user behaviors. However, this increased relevance comes at a cost, as deep learning recommenders can be difficult to train and expensive to serve at scale. 

Recommendations AI helps organizations take advantage of serving these deep learning models and handles the MLOps required to serve these models globally with low latency. Models are automatically retrained daily and tuned quarterly to capture changes in customer behavior, product assortment, pricing, and promotions. Newly trained models follow a resilient CI/CD routine which validates they are fit to serve and promotes them to production without service interruption. The models achieve low serving latency by using a scalable approximate nearest neighbors (ANN) service for efficient item retrieval at inference time. And, to maintain consistency between online and offline tasks, a scalable feature store is used, preventing common production challenges such as data leakage and training-serving skew.  

What’s required to get started?

To get started with Recommendations AI we first need to ingest product and user data into the API:

• Import product catalog: For large product catalog updates, ingest catalog items in bulk using the catalogItems.import method. Frequent catalog updates can be schedule with Google Merchant Center or BigQuery
• Record user events: User events track actions such as clicking on a product, adding items to cart, or even purchasing an item. These events need to be ingested in real time to reflect the latest user behavior and then joined to items imported in the product catalog 
• Import historical user events: The models need sufficient training data before they can provide accurate predictions. The recommended user event data requirements are different across model types (learn more here)

Once the data requirements are met, we are able to create one or multiple models to serve recommendations: 

• Determine your recommendation types and placements:  The location of the recommendation panel and the objective for that panel impact model training and tuning. Review the available recommendations types, optimization objectives, and other model tuning options to determine the best options for your business objectives.
• Create model(s): Initial model training and tuning can take 2-5 days depending on the number of user events and size of the product catalog 
• Create serving configurations and preview recommendations: After the model is activated, create serving configurations and preview the recommendations to ensure your setup is functioning as expected before serving to production traffic

Once models are ready to serve, consider setting up A/B experiments to understand how newly trained models impact your customer experience before serving them to 100% of your traffic. In the Recommendations AI console, see the Monitoring & Analytics  page for summary and placement-specific metrics (e.g., recommender-engaged revenue, click-through-rate, conversion rate, and more).

Why use this approach and who is it a good fit for?  

Recommendations AI is a great way to engage customers and grow your online presence through personalization. It’s used by teams who lack technical experience with production recommendation systems, as well as customers who have this technical depth but want to allocate their team’s effort towards other priorities and challenges. No matter your team’s technical experience or bandwidth, you can expect several benefits with Recommendations AI: 

• Fully managed service: no need to preprocess data, train or hypertune machine learning models, load balance or manually provision you infrastructure – this is all taken care of for you. The recommendation API also provides a user-friendly console to monitor performance over time. 
• State-of-the-art AI: take advantage of the same modeling techniques used to serve recommendations across Google Ads, Google Search, and YouTube. These models excel in scenarios with long-tail products and cold-starts users and items
• Deliver at any touchpoint: serve high-quality recommendations to both first-time users and loyal customers anywhere in their journey via web, mobile, email, and more
• Deliver globally: serve recommendations in any language anywhere in the world at low-latency with a fully automated global serving infrastructure
• Your data, your models: Your data and models are yours. They’ll never be used for any other Google product nor shown to any other Google customer

For users looking to leverage state of the art AI to fuel their recommendation systems but need an existing solution to get up and running more quickly, Recommendations AI is the right solution for you. 

What are Two Tower encoders and how do they work?

As a reminder, in recommendation system design, our objective is to surface the most relevant set of items for a given user or set of users. The items are usually referred to as the candidate(s) where we might include information about the items such as the title or description of the item, other metadata about the item like language, number of views, or even clicks on the item over time. User(s) are often represented in the form of a query to a recommendation system where we might provide details about the user such as the location of the user, preferred languages, and what they have searched for in the past.   

Let’s start with a common example. Imagine that you are creating a movie recommendation system. The input candidates for such a system would be thousands of movies and the query set can consist of millions of viewers. The goal of the retrieval stage is to select a smaller subset of movies(candidates) for each user and then score and rank order them before presenting the final recommended list to the query/user.

The retrieval stage is able to refine our list of candidates by encoding both the candidate and the query data so they share the same embedding space. A good embedding space will place candidates which are similar to one another closer together and dissimilar items/queries farther apart in the embedding space.

Once we have a database of query and candidate embeddings we can then use an approximate nearest neighbor search method to then generate a list of final “like” candidates, i.e. find a certain number of nearest neighbors for a given query/user and surface final recommendations.

What’s required to get started?

At the most basic level, in order to train a two-tower model you need the following inputs:

• Training Data: Training data is created by combining your query/user data with data about the candidates/items. The data must include matched pairs, cases where both user and item information is available. Data in the training set can include many formats from text, numeric data, or even images. 
• Input Schema: The input schema describes the schema of the combined training data along with any specific feature configurations.

Several services within Vertex AI have come available that complement the existing Two-Tower built-in algorithm and can be leveraged in your execution: 

• Nearest Neighbor (ANN) Service: Vertex AI Matching Engine and  ScANN provide a high-scale and low-latency Approximate Nearest Neighbor (ANN) service so you can more easily identify similar embeddings.
• Hyperparameter Tuning Service: A hyperparameter tuning service such as Vizier can help you identify the optimal hyperparameters such as the number of hidden layers, the size of the hidden layers, and the learning rate in fewer trials. 
• Hardware Accelerators: Specialized hardware, such as GPUs or TPUs, can be valuable in your recommendation system to help accelerate experiments and improve the speed of training cycles. 

Why use this approach and who is it a good fit for?  

The Two-Tower built-in algorithm can be considered the “custom sports car” of recommendation systems and comes with several benefits: 

• Greater Control: While Recommendations AI uses the two-tower architecture as one of the available architectures it doesn’t provide granular control or visibility into model training, example generation, and model validation details. In comparison, the Two-Tower built in algorithm provides a more customizable approach as you are training a model directly in a notebook environment. 

• More Feature Options: The Two Tower approach can handle additional contextual signals ranging from text to images. 
• Cold Start Cases: Leveraging a rich set of features not only enhances performance but also allows the candidate generation to work for new users or new candidates.

While the Two-Tower built in algorithm is an excellent and best-in class solution for deriving   recommendations, it does come with additional considerations and potential drawbacks depending upon the use case.

• Technical ML Expertise Required: Two tower encoders are not a “plug and play” solution like the other approaches mentioned above. In order to effectively leverage this approach, appropriate coding and ML expertise is required. 
• Speed to Insight: Building out a custom solution via two-tower encoders may require additional time as the solution is not pre-built for the user.   

For users looking for greater control, increased flexibility, and have the technical chops to easily work within a managed notebook environment – the two-tower built in algorithm is the right solution for them. 

What’s next?

In this article, we explored three common methods for building recommendation systems on Google Cloud Platform. As you can see thus far, there are alot of considerations to take into account before choosing a final approach. In an effort to help you align more quickly we have distilled the decision criteria down to a few simple steps (see below for more details).

In the next installments of this series, we will dive more deeply into each method, explore how hardware accelerators can play a key role in recommendation system design, and discuss how recommendation systems may be leveraged in key verticals. Stay tuned for future posts in our recommendation systems series. Thank you for reading! Have a question or want to chat? Find authors here – R.E. [Twitter | LinkedIn], Jordan [LinkedIn], and Vaibhav [LinkedIn].

Acknowledgements

Special thanks to Pallav Mehta, Henry Tappen,Abhinav Khushraj, and Nicholas Edelman for helping to review this post. 

References

• Announcing ScaNN: Efficient Vector Similarity Search
• Factorization Machines for Item Recommendation with Implicit Feedback Data

Every company these days is becoming a data company whether they know it or not. This results in preparing the company to create an ecosystem for data processing. Traditionally, organisations’ data ecosystems consisted of point solutions that used to provide data services. For example, one of the most common questions we get from customers is, “Do I need a data lake, or should I consider a data warehouse? Do you recommend I consider both?” Traditionally, these two architectures have been viewed as separate systems, applicable to specific data types and user skill sets. Increasingly,  we are seeing a blurring of lines between data warehouses and data lakes, which provide customers with an opportunity to create a more comprehensive platform that gives them the best of both worlds.

What if we don’t need to compromise? And we create an end-to-end solution compromising the entire data management and processing stages, from data collection to serving. Thus, Data platforms are typically used to store vast amounts of data in varying formats and doing so without compromising on latency. At the same time, providing a platform for all users throughout the data lifecycle. 

Emerging Trends

There are data solutions and architectures we’re seeing and anticipating. Emerging concepts include data lakehouses, data meshes, and data vaults. Some are not new and have been around in different shapes and formats, however, all of them work naturally within a Google Cloud environment. Lets look into both ends of the spectrum of enabling data and enabling teams. 

Data mesh facilitates a decentralized approach to data ownership, allowing individual lines of business to publish and subscribe to data in a standardized manner, instead of forcing data access and stewardship through a single, centralized team. On the other hand, a data lake house brings raw and processed data closer together, allowing for a more streamlined and centralized repository of data needed throughout the organization. Processing can be done in transit via ELT, reducing the need to copy datasets across systems. This allows for easier data exploration and easier governance. Data vault is designed to separate data driven and model driven parts, so the way data is integrated in the raw vault enables parallel loading so that large implementations can scale out easily.

In Google Cloud, there is no need to keep them separate. In fact, with interoperability among our portfolio of data analytics products, you can easily provide access to data residing in different places, effectively bringing your data lake and data warehouse together on a single platform.

Let’s look at some of the technological innovations that make this reality. BigQuery’s storage API allows treating a data warehouse like a data lake, letting you access the data residing in BigQuery. For example, you can use Spark to access data residing in the data warehouse without it affecting performance of any other jobs accessing it. This is all made possible by the underlying architecture, which separates compute and storage. 

We continue to offer specialized products and solutions around data lake and data warehouse functionality but over time we expect to see a significant enough convergence of the two systems that the terminology will change. At Google Cloud, we consider this combination an “analytics data platform”.

Tactical or Strategical

Key differentiators of Google Cloud’s data analytics platform are being open, intelligent, flexible, and tightly integrated. There are many technologies in the market which provide tactical solutions that may feel comfortable and familiar. However, this can be a rather short-term approach that simply lifts and shifts a siloed solution into the cloud. In contrast, an analytics data platform built on Google Cloud provides modern data warehousing and data lake capabilities with close integration to our AI Platform. It also provides built-in streaming, ML, and geospatial capabilities and an in-memory solution for BI use cases. Depending on your organizational data needs, Google Cloud has the set of products, tools, and services to create the right data platform for you. 

To become a truly data-driven organization, the first step is to design and implement an analytics data platform that meets your technical and business needs. Whether you want to empower teams to own, publish, and share their data across the organization, or you want to create a streamlined store of raw and processed data for easier discovery, there is a solution that best meets the needs of your company.

To learn more about the elements of a unified analytics data platform built on Google Cloud, and the differences in platform architectures and organizational structures, read our Unified Analytics Platform paper.

Japan is home to millions of vending machines installed on streets and in buildings, sports stadiums and other facilities. Vending machine owners and operators, including beverage manufacturers, stock these machines with different product combinations depending on location and demand. For example, they primarily display coffee and energy drinks in machines placed in offices and sports drinks and mineral water in machines at sports facilities. The combinations also vary by season: for example, owners and operators may display cold beverages in summer and hot beverages in winter. 

Traditionally, vending machine operators have relied on the intuition and experience of sales managers to determine the optimum product mix for each vending machine. However, in recent years, manufacturers such as Coca-Cola Bottlers Japan (CCBJ) have turned to data to analyze and make strategic decisions about when and where to locate products in machines.

CCBJ is the number one Coca-Cola bottler in Asia and vending machines comprise the bulk of its business. The organization operates about 700,000 machines across Tokyo, Osaka, Kyoto, and 35 prefectures. Minori Matsuda, Google Developer Expert and also Data Science Manager at CCBJ, says “The billions of data records collected from 700,000 physical devices are a great asset and a treasure trove we can take advantage of.”

Minori points out that when considering the mix of products in vending machines in sporting facilities, the managers naturally assume sports drinks would generally sell well. However, analysis of purchase data – including hot drinks and hot drinks plus sports drinks – found many parents purchased sweet drinks such as milk tea when they attended games or sessions involving their children.  “Analyzing data gives us new discoveries and, by using catchy storytelling techniques from exploratory data analysis, we are instilling a data culture within our company,” he says. “It’s worth creating by looking at facts rather than making assumptions!”

Minori believes that to analyze the vast amount of data collected from more than 700,000 vending machines, the business needs a powerful analytical platform. However, until recently, CCBJ had to extract data for analysis from its core systems, load this data into a warehouse it created and perform the required analyses.  The billions of records of data generated across the fleet – including transaction data – exposed some challenges for traditional analysis platforms. They could not efficiently process data at a considerable scale: it could take a day to return results and required extensive maintenance due to the size.

CCBJ considered building a machine learning (ML) platform as a layer on top of existing systems in August 2020 and opted for Google Cloud the following month.  “I feel that Google Cloud has an edge in all products and is very well thought out,“ says Minori, noting the scalability and cost of the platform allow the business to take a ‘trial and error’ approach to achieve the best outcomes from ML. Google Cloud also delivered the required visibility and flexibility to help the business deliver change every day against key performance indicators. 

MLOps platform streamlines ML pipeline development

CCBJ built its analysis platform using Vertex AI (formerly AI Platform) centered on a BigQuery analytics data warehouse, and partly using AutoML for tabular data. “We have created a prediction model of where to place vending machines, what products are lined up in the machines and at what price, how much they will sell, and implemented a mechanism that can be analyzed on a map,” says Minori, adding that building the platform with Google Cloud was not difficult. “We were able to realize it in a short period of time with a sense of speed, from platform examination to introduction, prediction model training, on-site proof of concept to rollout.”

The new data analytics platform of CCBJ consists of the following parts:

Data Sources

• The data collected from the vending machines are all stored on BigQuery.

Data Discovery and Feature Engineering

• Minori and other data scientists at CCBJ are using Vertex Notebooks, where they access the data on BigQuery by executing SQL queries directly from the Notebooks. This environment is used for the data discovery process and feature engineering. 

ML Training

• For ML training, CCBJ uses AutoML for Tabular data, Custom model training on Vertex AI, and BigQuery ML. AutoML gives model performance with AUC curves and also feature importance graphs.

ML Prediction and Serving

• For ML prediction, CCBJ uses Online Prediction for AutoML models and Online Prediction for custom models for real-time prediction when the salesperson finds the interesting point
• Batch Prediction is used for generating a large prediction map that covers the whole country
• The prediction results are distributed to sales managers’ tablets

CCBJ started constructing the platform in September 2020, and completed it within a month. The business has conducted proofs of concept at its base in Kyoto since February 2021, and since April, has rolled out the platform to sales managers in 35 prefectures in one metropolitan area. “Data analysis is built into the day-to-day routines of sales managers with 100% utilization,” says Minori. “They can utilize the prediction results on tablets that were able to achieve pretty high accuracy from the start.”

The hardest part was the education of sales managers in the field; having them understand the reasoning behind the ML prediction results for particular outcomes, so they could be convinced to make use of the results. “For example, regarding a new installation location predicted by the model, it seemed that there was no effective information for installation from the map information, but when I actually went there, there was a motorcycle shop and it was a place where young people who like motorcycles gathered,” says Minori. “Or there is a small meeting place where the elderly in the neighborhood are active. 

“In many cases, new discoveries that cannot be understood from map information alone can be derived from the data.”

Minori also points to a phenomenon whereby humans pursued and confirmed factors inferred by the model – meaning that once they experienced analysis and it worked effectively, they asked why the same type of analysis or prediction could not be undertaken next time. The resulting cycle of more inquiries generated, more information gathered and more data captured for analysis meant the accuracy of results was improved.

Minori describes Vertex AI as having a number of strengths in helping CCBJ build a ML data analysis platform. “One of the major merits of Vertex AI was that we were able to realize MLOps that streamlines the entire development life cycle from construction of the ML pipeline to its execution,” he says.

With near real-time data analysis through Google Cloud, CCBJ teams can spend time developing strategies rather than waiting for data requested from the IT systems department. Exploratory data analysis is also considerably easier as repeated trial and error has greatly improved the accuracy of analyses. Before we used Machine Learning, most machine placement processes were done by human senses, by looking at a map to find the suggestion points. By using Machine Learning to generate a massive number of placement point suggestions, the efficiency of routing of salespeople has been dramatically improved. 

In the future, CCBJ aims to automate the continuous training pipeline with Vertex AI. “CCBJ is a tech company that operates in the food industry,” says Minori. With the organization operating a vending machine network of 700,000 units, it would like to create new businesses based on utilization and analyzing data. Some of these businesses may be based on Sustainable Development Goals (SDGs) initiatives such as the utilization of recycled PET bottles, measures to prevent food loss and ways of using vending machines to contribute to local communities, which we have been working on for some time. It would be interesting if we could collaborate with Google Cloud on these in the future.”