When it comes to the prediction of continuous variables, the first thing that comes to our mind is always the regression model. For instance, linear regression is the most commonly used regression model, and it has the benefits of simple implementation and high interpretability. On the other hand, random forest regression can handle missing data and is adaptive to interactions and nonlinearity. While these algorithms all work well for continuous target variables in different scenarios analytically, they provide less information on the predicted numbers’ confidence level, especially in real-world applications.

In this article, we will explore an unconventional framework to predict continuous variables with given confidence scores. Instead of framing the prediction as a regression problem, we twist the problem into a classification problem. This framework also allows us to have more visibility on the predicted results and can be adjustable to different confidence levels. The article will use revenue estimation as an example. Given a variety of business attributes for many businesses, we will illustrate how we can predict the revenue for each business given a specific confidence level.

Many impactful business problems need to be translated into corresponding machine learning tasks. For example, we must be able to recognize fraudulent credit card transactions to prevent loss for both our merchants and consumers. Credit card fraud detection is often translated to a classification task in machine learning.

Recommending the right financial products to the right customers is another important problem. One way to tackle product recommendation is by computing a propensity score (i.e. the likelihood of adopting a product for a user). Effective product recommendations and promotional strategies are often built based on product propensity scores. Propensity modeling, which determines the best product-customer pairs, is often translated into a classification task.

When our customers make purchases with credit or debit cards in their PayPal wallet, we send the payment details through card networks to card issuing banks for payment authorization. During this process, payments can be declined by the issuing banks for various reasons, such as the card: exceeding its credit limit, not having enough balance, being reported as stolen, or outdated information(see here for more information regarding issuer declines causes).

Apart from these reasons, issuers also consider many other factors when deciding to approve or decline a payment request, such as: what type of card is being used for the payment? What merchandise is being purchased? What time is the payment request submitted? What is the card’s approval and decline history from issuers?

Text resizing is the ability for the user to adjust their preferred font size in their device’s operating system preferences. A variety of users make this adjustment to improve their ability to read small text on a small screen. Once this setting is adjusted, the platform and any apps that have built-in support for this feature will resize displayed text. Check out these examples of how to adjust these text and font resizing settings in Android and iOS platforms. Pro tip: If you don’t already use these settings, grab your phone and try it right now. As we all know, it’s good to put yourself in your customer’s shoes. Plus it’ll be good practice because most of us will find the feature helpful eventually.

In this article I’d like to share my experiences on how one can improve the observability of services through the use of the Consul Connect Service Mesh and later I’ll explain my justification for using it compared to other alternatives. Just for a recap, a service mesh allows you to route HTTP traffic through proxies that are deployed as sidecar containers to a service. These proxies can then expose metrics, logs and traces that allow you to gain rich visibility into the system thus making it more efficient to debug applications quickly and easily. Consul Connect, in addition provides service-to-service authorization and encryption using mutual TLS and uses Envoy as the data plane for the data transmission. In terms of observability, I will cover basic Layer 7 metrics, access logs and distributed tracing using Prometheus, Grafana, AWS CloudWatch Logs and AWS X-Ray.

Spark, Hive and HDFS (Hadoop Distributed File Systems) ecosystems are online analytical processing (OLAP)-oriented technologies. They are designed to process huge amounts of data with full scans. From time to time, users want to use the same data for more ad-hoc oriented tasks:

• Multi-row load— explore small sets (typically 1%) of the data by specific IDs (not random).
• Single-row fetch — for example, building a serving layer to fetch a specific row upon a REST-API request.

These kinds of tasks are traditionally solved using dedicated storage and technology stacks (HBase, Cassandra, etc.) which require data duplication and add significant operational costs.

In this post, we describe our journey for solving this challenge by using only Spark and HDFS. We will start by introducing an example use case, generalize and define the requirements, suggest some optional solutions, and finally dive into our final solution.

Re-ordering table columns to avoid wasting space for data type alignment.

Braintree Payments uses PostgreSQL as its primary datastore. We rely heavily on the data safety and consistency guarantees a traditional relational database offers us, but these guarantees come with certain operational difficulties. To make things even more interesting, we allow zero scheduled functional downtime for our main payments processing services.

Imagine a project failed, and start from there.

We recently added multi-tenancy support in our Mobile Software Development Kits (SDKs) and learned a lot during that process. I want to share my experience with the technology community as a quick start guide for developers who are interested in making their SDKs multi-tenant friendly.

This article focuses on Android and is based on the architecture, design, and patterns we used in Android Assistant SDK for adding multi-tenancy support.

减少洲际文件传输的延时

想象一下，你正在与全球各地的人进行重要的视频通话，但是当你开始说话的时候，你的队友在你的线路上只听到一个破碎的机器人声音。遇到这种情况，对你和电话另一端的人来说是多么令人沮丧？

这些抖动或延迟问题对我们来说并不陌生。它们甚至可能与底层软件无关，而更多的是与网络的设置方式有关。

在PayPal，由于我们总是希望以最高的质量提供最好的客户体验，我们通过使用CDN和边缘计算优化网络设置，解决了相当多的延迟问题。我们现在在全球各地设置了边缘服务器，以低延迟的方式提供网页。

这篇文章涵盖了PayPal DropZone，一个安全的文件传输平台，如何使用现有的PayPal Edge基础设施改善文件传输的延迟和可用性。

PayPal和其他大公司一样，有许多数据中心、地区和区域，有不同的安全级别和限制来保护数据。这使得数据移动成为一项不容易的任务，对维护安全和数据保护合规性提出了很高的要求。

企业数据平台(EDP)的数据移动平台(DMP)团队需要建立一个快速可靠的数据移动通道，提供给EDP内外的客户。这个通道需要是安全的，完全符合InfoSec的要求。去年，我们开始了构建下一代数据移动平台的旅程。这篇博客专门介绍平台的安全方面。

我们仍然看到了中国企业进一步实现数字化转型的强大动力。

A few years ago, when NoSQL was trending, like every other team, our team was also enthusiastic about the new and exciting stuff; we were planning to change the database in one of the applications. But when we got into the finer details of the implementation, we remembered what a wise man once said, “the devil is in the details,” and eventually, we realized that NoSQL is not a silver bullet to fix all problems, and the answer to NoSQL VS RDMS was: “it depends.” Similarly, in the last year, concurrency libraries like RxJava and Spring Reactor were trending with enthusiastic statements, like the asynchronous, non-blocking approach is the way to go, etc. In order to not make the same mistake again, we tried to evaluate how concurrency frameworks like ExecutorService, RxJava, Disruptor, and Akka differ from one another and how to identify the right use case for respective frameworks.

At PayPal, we take security seriously. Since the client-secret in the API world is akin to your password in the web world, it is a…