This is the second part of our three-part blog post series (see the first part here), which deals with incremental data updates. In our scenario we assume that we acquire small batches of data updates using some kind of web scraping mechanism. We will not deal with the details of that mechanism, as it is beyond the scope of this …
Graph analysis of Stack Overflow tags with Oracle PGX – Part 1: Data Engineering
Intoduction Oracle Parallel Graph Analytics (PGX) is a toolkit for graph analysis, both for running algorithms such as PageRank and for performing SQL-like pattern-matching against graphs. Extreme performance is offered through algorithm parallelization, and graphs can be loaded from a variety of sources such as flat files, SQL and NoSQL databases etc. So, in order to get a deeper feeling, …
Enabling the Green-Marl compiler for Parallel Graph Analytics in Oracle Big Data Lite VM
Recently, I began working with Parallel Graph Analytics (PGX) on my Oracle Big Data Lite (BDL) VM version 4.7.0.1. I was especially intrigued and curious about the capabilities of a PGX component called Green-Marl (GM), a domain-specific language specially designed for graph data analysis. It was stated to extend PGX’s capabilities and “implement algorithms with no limit”. Especially the last argument …
Streaming data from Raspberry Pi to Oracle NoSQL via Node-RED
Starting from version 4.2, Oracle NoSQL now offers drivers for Node.js and Python, in addition to the existing ones for Java, C, and C++; this is good news for data science people, like myself, since we are normally not accustomed to code in Java or C/C++. So, I thought to build a short demo project, putting into test both the …
sparklyr: a test drive on YARN
sparklyr is a new R front-end for Apache Spark, developed by the good people at RStudio. It offers much more functionality compared to the existing SparkR interface by Databricks, allowing both dplyr-based data transformations, as well as access to the machine learning libraries of both Spark and H2O Sparkling Water. Moreover, the latest RStudio IDE v1.0 now offers native support …
Nonlinear regression using Spark – Part 2: sum-of-squares objective functions
This post is the second one in a series that discusses algorithmic and implementation issues about nonlinear regression using Spark. In the previous post we identified a small window for contribution into Spark MLlib by adding methods for nonlinear regression, starting with the definition and implementation of a general nonlinear model. We remind the reader that regression is essentially an …
Classification in Spark 2.0: “Input validation failed” and other wondrous tales
Spark 2.0 has been released since last July but, despite the numerous improvements and new features, several annoyances still remain and can cause headaches, especially in the Spark machine learning APIs. Today we’ll have a look at some of them, inspired by a recent answer of mine in a Stack Overflow question (the question was about Spark 1.6 but, as …
How to use SparkR in Cloudera Hadoop
Suppose you are an avid R user, and you would like to use SparkR in Cloudera Hadoop; unfortunately, as of the latest CDH version (5.7), SparkR is still not supported (and, according to a recent discussion in the Cloudera forums, we shouldn’t expect this to happen anytime soon). Is there anything you can do? Well, indeed there is. In this …
Bulk load data to HBase in Oracle Big Data Appliance
I ran into an issue recently, while trying to bulk load some data to HBase in Oracle Big Data Appliance. Following is a reproducible description and solution using the current version of Oracle Big Data Lite VM (4.4.0). Enabling HBase in Oracle Big Data Lite VM (Feel free to skip this section if you do not use Oracle Big Data …
Querying Big Data SQL tables with Oracle R Enterprise
I was wondering recently if I could use Oracle R Enterprise (ORE) to query Big Data SQL tables (i.e. Oracle Database external tables based on HDFS or Hive data), since I have never seen such a combination mentioned in the relevant Oracle documentation and white papers. I am happy to announce that the answer is an unconditional yes. In this …