5 Best Linux Business Intelligence SuitesBy Carla Schroder | Published on: 03-May-12
Even a small business generates giant quantities of data, and a good business intelligence suite helps you analyze and make sense of it all. When you have an accurate picture of where you are, you'll see where you can go, and any of these excellent Linux-based small business intelligence suites will take you there.
Business intelligence (BI) software isn't magical, so don't go throwing a ton of money into it before thinking about what you want to accomplish with it. All of these suites are modular, so you can start small and add more modules as you get familiar with the software and figure out what sort of information you want.
Measuring actual results is a surprisingly neglected task in a lot of businesses, and is often an eye-opener -- is the new product really paying off, and what markets is it doing best in? What are your real costs and rewards in supporting a branch office, or in clinging to old equipment, or in staying open on Sundays and holidays? BI can identify bottlenecks and inefficiencies, and show you what's working and what isn't.
BI software requires some expertise to install, set up, and maintain. When you're shopping for BI software you need to consider:
- Do you have in-house expertise, or do you need to hire help?
- Integration with your existing databases and data sources
- Integration with other software (for example CRM/ERP apps like SugarCRM, Salesforce, Sage)
- Open data formats, support for multiple data formats
- Licensing mode: per server, per user, or something else? Per user licenses for a server product get expensive fast, and they make no sense. It's like charging per person for tap water.
- If you hire custom engineering services, how quickly will they have you up and running? All of the BI vendors in this roundup promise speedy service, as quickly as three to eight weeks.
All of these BI suites have more in common than they have differences: they run on proven open source technologies such as Apache Tomcat and JBoss, AJAX, Java, and various scripting languages, they support multiple databases and data sources, and multiple Web browsers. Competition is pretty fierce and you have good choices, so it may come down to which vendor understands your needs the best, and who gives you the best deal.
Figure 1: Pentaho Data Mining and Predictive Analytics. Image courtesy Pentaho
BI depends on well-organized databases, so you may have some cleanup and organizing to do before you get to play with shiny new BI software. When it's well-setup then non-technical users should be able to slice and dice data however they need to. It isn't a quick fix, but a long-term investment that, used well, pays for itself many times over.
Pricing BI Software
Pricing takes a bit of work to nail down; it pays to talk to the nice salespeople and let them help you sort it all out. All of these BI suites have free software downloads, and free community support. All of them offer multiple modules, custom support and engineering services, and training and documentation. So the answer to "how much will it cost" is always "it depends."
For example, Jaspersoft sells self-support packages for its free community edition that range from $99 to $999 per year for 1,800+ pages of product documentation, access to the knowledge base, and a number of Web-based support incidents. SpagoBI offers three levels of email support that range from about $1,300 to $6,500 per year.
Online training courses range from free to a few hundred dollars and even to $1,200+ for live online sessions.
Software licenses and custom engineering are especially "it depends." Your total first-year costs can easily hit $10,000 for a small shop. Presumably you'll spend less on support and training after your first year. Still, that's considerably less than the traditional proprietary offerings from SAP, IBM Cognos, SAS and other old-time proprietary BI vendors.
Jaspersoft claims to be the world's most-used BI software. JasperReports is everywhere; this is the popular core Java-based reporting engine used in many BI suites to pull data from multiple sources, and output it into multiple formats such as Web pages, PDF, XML, spreadsheet, DOC, and ODT.
JasperReports powers the Jaspersoft iReport Designer, which gives non-technical users the capability to create complex, sophisticated reports. In fact Jaspersoft's strength is its excellent tools for end-users to slice, dice, and present data without having to be ace database gurus.
Talend is the Jaspersoft ETL (extract, transform, load) module, which is also used in numerous BI suites. This is the middleware that integrates data from multiple sources. Nobody has a single harmonious squeaky-clean database; it's always ad-hoc and scattered, so an ETL layer is essential.
Jaspersoft is available as on-premises software, and of course as a hosted cloud service (which everyone, it seems, offers these days) and as a slick mobile BI interface for viewing and building reports. There is a free community edition, and multiple professional editions with different features and support levels.
Spago is a pure open-source BI platform. All of the software is open source and free of cost, as opposed to the more common model of a limited-feature community edition, and then advanced features that cost money. Support services, custom development, and recorded training courses are available for purchase. Spago is based in Italy, and the documentation is published in French and English.
Spago provides a number of different prefab domains, which are SpagoBI tailored for different use cases. For example, a retail business could use Real-time BI to track sales trends hourly, and make sure that hot items are always stocked. The Mobile BI works in concert with Real-time BI, so that managers can track trends from any location. GeoBI builds maps showing location-based analysis of patterns and trends. Spago also offers custom services and will build your BI however you need.
Pentaho is the other big open source BI vendor. Pentaho, like Jaspersoft, has excellent dashboards and report builders for non-technical users. Pentaho incorporates a lot of nice charting features like scatterplots, geo-mapping, bubble charts, lasso filtering, heat grids, and multiple ways to highlight data points.
Pentaho has some of the best built-in support for, well, everything: Hadoop clustering, all kinds of databases, cloud vendors, all kinds of middleware; it's also very supportive of ISVs (independent service vendors), so if you are an ISV you might look at being a Pentaho integrator.
Pentaho has invested a lot of resources into making Pentaho embeddable into nearly any environment, and it has extensive APIs (application programming interfaces) and Web services support, with the goal of making it a seamless component in any datacenter.
OpenI has the most eye-catching name in this roundup, and a different approach than the others. OpenI started out as a complete BI platform, but this led to a lot of duplicated effort with other open source BI products. So now OpenI is is a collection of plugins for Jasper and Pentaho, the two leading open source BI suites.
OpenI aims to provide functionality not found in Jasper and Pentaho, such as better user interfaces and better tools for exploring OLAP (Online Analytic Processing) cube data. What is this OLAP cube thingy, you ask? An excellent question. In a nutshell it's a multi-dimensional representation of your data that is designed to answer any question you might have, and to answer it quickly. Figure 2 should help you visualize it.
Figure 2: OLAP lets you slice and dice complex data almost any you want, quickly. (Image courtesy Wikipedia, author Infopedian, CC BY-SA 3.0)
OLAP lets you make arbitrary, unstructured queries so you can parse your data any way that makes sense to you. For example, sales data might be aggregated per salesperson, per district, per different time periods, per different products or product lines, per wholesale costs, tax data...it's a rare BI suite that doesn't have good OLAP support.
OpenI designs dashboards, predictive models, interactive reporting, and custom design services, claiming "Data to Insights in 72 Hours."
Actuate is built on BIRT, the open source Business Intelligence and Reporting Tools engine. Actuate's emphasis is on presentation, and you can make some pretty flashy animated presentations that incorporate the latest Web 2.0 technologies.
The BIRT Viewer is a read-only presentation that can't be altered by the people viewing it. The BIRT Interactive Viewer allows viewers to modify the content. BIRT Designer Pro is aimed at software developers for building Web reports and dashboards, but it's also a nice tool for tech-savvy managers. Actuate includes a full complement of necessary BI tools such as a mobile interface, data integration, and powerful analytics.
Actuate also emphasizes consistency: one server, a consistent look and feel, and a consistent user experience. Actuate has both free community-supported open source downloads, and an assortment of commercial options.
Carla Schroder is the author of The Book of Audacity, Linux Cookbook, Linux Networking Cookbook, and hundreds of Linux how-to articles. She's the former managing editor of Linux Planet and Linux Today.
|Do you have a comment or question about this article or other small business topics in general? Speak out in the SmallBusinessComputing.com Forums. Join the discussion today!|