HDInsight Hadoop Cluster is not Responding Issue

31 Mar
March 31, 2013

This is a quick post to describe a solution to a problem I was having with my local installation of the HDInsight Hadoop cluster.

Read more →

Rehashing SQL Server Hashing Algorithms for Large Text Fields

23 Mar
March 23, 2013

Hashing can be a very useful technique when dealing with the storage and look up of large text fields (say a table of URLs or Search Keywords), these fields will incur high resource utilization on any database engine if used directly in DML statements, in which they are either filtered by or aggregated on. Any index built on these fields is costly to maintain, if it is at all possible given that SQL Server limits index size to 900 bytes.

Using hashing functions we can facilitate the handling of large textual data in the relational engine, leading to improved performance when these fields are being compared to satisfy a query, hashing can also be used to build unique and non-unique indexes that are easier to manage than directly using the text fields in the index definition. In this post we will discuss a few options for hashing large text data using functions native within SQL Server, as well as provide other external  hashing algorithms that we can integrate into Microsoft’s SQL Server (or any RDBMS for that matter) that might provide a better practical performance. 

Read more →

SSIS vs Hadoop – a Mapping Performance Showdown

28 Feb
February 28, 2013

On one corner we have Hadoop, a massively distributed JVM-based data processing engine with a Map & Reduce API and a proven track record in handling huge data-sets. On the other corner we have SSIS, a natively non-distributed ETL engine part of the SQL Server family tool-set with .NET code extensibility features and a drag and drop UI (for the most part anyway). Two sweet technologies, probably shouldn’t be compared to each other but we’re doing it anyway, pitted head to head against a data mapping task to the death (or at least to the recycling of my test VMs)… Now FIGHT!

Read more →

Architectures for Running SQL Server Analysis Service (SSAS) on Data in Hadoop Hive

25 Feb
February 25, 2013

Recently I have been involved in researching and building a low-latency high-data-volume OLAP environment for a social entity and interaction analysis platform, the perfect mixture of concepts such as Big Data collection and processing,  large-scale Network Analysis, Natural Language Processing (NLP) and a highly scaled-out OLAP environment for end users to explore and discover data (essentially a Self-Service and Exploratory BI layer).

It is by all means not an easy mission to orchestrate all the technologies that back those concepts, particularly if you are interested in using the optimum solution for the problem at hand, for example Big Data might be better handled by a Hadoop layer, but Hadoop or Hive (at least on their own) are not geared up to respond to OLAP queries, which are real-time by nature, and even if they were, your end-user needs familiar tools and interfaces to analyse and study this data, which is where SQL Server Analysis Service and the whole Microsoft BI stack might come in and offer great integration with already existing business applications (such as Office or SharePoint).

This post discusses a few architectural approaches to exposing a Hadoop layer through a SQL Server Analysis Service (SSAS) interface, with references to data-latency, redundancy and over-all performance.

Read more →

Running Highcharts within SSRS (or any JS Graph Library)

22 Jan
January 22, 2013

In a previous post I described how to convert an SSRS graph into a Highcharts graph by consuming the XML output of the report from the SSRS Web Service and converting that to an input for a Highcharts graph.

That article seemed to be very popular (in fact was the top most popular for a while), so I decided to take this concept a step further, In this article I will show you how, using JavaScript injection into SSRS reports, you can display a Highcharts graph from within SSRS itself (just like any other SSRS report) when the SSRS report is rendered into HTML.

Read more →

10 Tips to Improve your Text Classification Algorithm Accuracy and Performance

21 Jan
January 21, 2013

In this article I discuss some methods you could adopt to improve the accuracy of your text classifier, I’ve taken a generalized approach so the recommendations here should really apply for most text classification problem you are dealing with, be it Sentiment Analysis, Topic Classification or any text based classifier. This is by no means a comprehensive list, but it should provide a nice introduction into the subject of text classification algorithm optimisation.

Read more →

Text Classification Threshold Performance Graph

20 Jan
January 20, 2013

One way to increase the accuracy of a classification algorithm is to allow the algorithm to return an “Unknown” value, particularly when the probability of what we are trying to classify is too low to simply belong in one class and the algorithm is essentially guessing an answer, leading to incorrect classification.

In this post I will try and explore a method for researching and implementing the “Unknown” result in your classifier based on the probability distribution results of a classification, the idea is to give you the tools to tweak the optimum thresholds that gives you the best accuracy, while maintaining acceptable level of over-all coverage of data.

Read more →

SQL Saturday in Edinburgh on June 2013

20 Jan
January 20, 2013

SQL Saturday is finally coming to Scotland with a session scheduled for Edinburgh in June  2013.

Pretty damn exciting if you’re into the whole SQL Server scene, and considering I live in Edinburgh you bet I’ll be there (which might sway your decision not to go). Last year I missed SQL Saturday in Dublin because I couldn’t sort out travel arrangements in time, which was such a disappointment considering everyone else on my team ended up going.

Read more →

Testing & Diagnosing a Text Classification Algorithm

19 Jan
January 19, 2013

To get something going with text (or any) classification algorithm is easy enough, all you need is an algorithm, such as Maximum Entropy or Naive Bayes, an implementation of each is available in many different flavors across various programming languages (I use NLTK on Python for text classification), and a bunch of already classified corpus data to train your algorithm on and that is it, you got yourself a basic classifier.

But the story rarely ends here, and to get any decent production-level performance or accuracy out of your classification algorithm, you’ll need to iteratively test your algorithm for optimum configuration, understand how different classes interact with each other, and diagnose any abnormality or irregularity you’re algorithm is experiencing.

In this post I hope to cover some basic mathematical tools for diagnosing and testing a classification algorithm, I will be taking a real life algorithm that I have worked as an example, and explore the various techniques we used to better understand how well it is performing, and when it is not performing, what is the underlying characteristic of this failure.

Read more →

Running Windows 8 on Raspberry Pi

18 Jan
January 18, 2013

… Essentially you can’t run Win8 on the Raspberry Pi.

In this article I’ll try and explore the reasons why the Raspberry Pi is unable to support Windows 8, as well as present some alternatives that might achieve a sub-set of the Windows OS functionality.

Read more →