featured, Photography

Long Exposure Shots using Military Shooting Techniques

A couple of weeks ago I went on a night photography tour around London with a few friends. I knew most if not all photos I will be taking would be long exposure shots. I faced a real challenge as I wasn’t armed with a tripod. I thought I would use ledges or park benches or something like that to rest my camera on for these shots. This only worked in cases where the position was right for the framing of the shot, but there was still some vibrations from the shutter clicks.

When holding the camera, I knew some basic techniques like holding my breath for a few seconds to get my body stable. This wasn’t enough even for exposure of 2 seconds. So I started researching about how to stabilise the body (mine and camera’s) for these shots.

I came across a great article written by an officer in the US Army. He details the shooting (firing) technique used for rifle shooting and how this can be applied to photography. His explanations are clear and are illustrated. There are “four fundamentals of marksmanship”: Body posture, Breathing control, Aiming and Sight, and Trigger (shutter) squeeze.

This is worth a read (link below) if you want some great long exposure shots without a tripod. Although I wouldn’t recommend a 30s exposure with this technique!

SOURCE: PentaxForums , image

BI, Technology

Dimensional Modelling

I’m sure you have come across terms like Facts, Measures, Dimensions, Attribute hierarchies, etc. We will define these terms and see how a dimensional model is built.

A GRAIN is the smallest piece of information that needs to be presented in the warehouse solution.It is the most atomic piece of information in the data warehouse. It is best to give an example: For retail systems, a grain can be the items per transaction per customer. This is the most exact information that the warehouse will be able to provide. In defining the grain we define the fact too.

A FACT or MEASURE is something of interest. Something that we need to measure to give us an indicator to base our business decisions on. It is the most atomic piece of information used for aggregations. For example, for an inventory system this would be the items in stock; so we count (aggregate) the items to give us number of items. The fact here is an individual item, we measure the number of items. Another example can be . The piece of information of interest is customer and thus we would aggregate on this to find he number of customers enrolled or number of customers who shop at a particular store.

A DIMENSION is the information that we want to group the facts by. In the case of the inventory system, we can group the items by the type of item; eg. food, furniture, electronics, clothes, etc. For the fact customers we group them by the type of membership they hold or their city of residence. A dimension holds information related to the facts we measure.

Every dimension has attributes. The attribute usually is a property of the grouping or a sub category. For the dimension of geography, possible attributes are: continent, country, state, city. This information is grouped and described as “geography”.

This should give you an idea of what the terms mean. Once you understand these terms, you are ready to translate the business requirements into a warehouse model. In the next post we will look at requirements gathering and how to get the business users to narrow down their high level objectives.

featured, Technology

Google Chrome (beta) for Android

Early his week Google released a beta version of their Chrome browser for Android 4.0 (Ice Cream Sandwich). You can download the beta from Android Market.

The native web browser in Android has shared the webkit with the desktop version of Chrome for a while now. The performance improvements are related to JavaScript execution and it feels well optimised for a mobile environment. The UI is very simple and similar to the desktop version. Some nice features like tab organisation are well implemented.

Instead of talking about the UI, here are some screenshots.

Try out Chrome and let me know what you think.

BI, Business

Data Warehousing – A Process Overview

Data warehouse is a very common term that you will come across in the BI field. A data warehouse is simply a data store in its most basic definition. This data store is used for analysis and reporting and not for operational use.

The data warehouse typically has 3 layers: staging, integration and access layers. The staging layer usually contains a raw data dump from an operational database. The integration layer is where the raw data is restructured and links the data in a meaningful manner. The access layer is used for presentation and is usually for business clients.

The data warehouse stores all the business data for a corporation. This can be subdivided and stored into smaller stores known as Data Marts. Data Marts are typically subdivided by business function or logical reporting categories. If a single large Data Warehouse is implemented then this is known as centralised approach. The data is centrally stored and is accessed by all business functions. Implementing multiple data marts is a decentralised approach; as there is no single store with information from all business functions.

It’s beginning to sound like the two approaches are very different but they have a similar purpose. In fact, Ralph Kimball and Bill Inmon are the authors behind the decentralised and centralised approaches respectively. Inmon defined the idea of centralised storage and a top down approach. A top down approach starts with consideration of all the requirements at an enterprise level without breakdown by business function. Kimball defined a bottom up approach leading to decentralised storage with multiple Marts. A bottom up approach starts with individual business functions and their specific requirements. After all the marts are built these are then linked together to deliver an enterprise-wide solution.

Even though these two approaches to building a DW are different both involve 3 common stages: Integration, Analysis and Reporting. Integration and Analysis both involve manipulation of data, whereas Reporting is more about presentation of the data. It is possible to perform further analysis in the reporting layer.

The tools used for development usually support these three layers. One of the industry leading tool sets is provided by Microsoft. The SQL Server with Business Intelligence Development Studio (BIDS) supports the development of the above mentioned layers. The SQL Server Integration Services (SSIS), SQL Server Analysis Services (SSAS) and SQL Server Reporting Services (SSRS) together form the MS BI stack.

SSIS is used to get data from an operational database and transform it into a dimensional or ER data store. SSAS allows building of a multi-dimensional Cube using the new DW model and perform calculations and aggregations. SSRS then allows generating reports from the Cube with some graphics.

In the next post we will take a closer look at the dimensional model (Kimball) for data warehousing and define some basic terms.