Welcome to WordPress. This is your first post. Edit or delete it, then start blogging!
I was watching the original Toy Story movie with my kids the other day. Even watching it today, it’s a surprising movie that still stands up after 15 years. For the uninitiated, the movie revolves around the misadventures of a toy cowboy named Woody. In the movie, Woody escapes with Buzz Lightyear from the clutches of Sid, a twisted neighborhood kid who tortures toys in various ways. All of this is in an attempt to return home, so Woody and Buzz wont get left behind when their owner moves.
One of the pivotal scenes is when Woody is in Sids lair, where he is surrounded by some of Sids twisted creations: the head of a pterodactyl on a dolls body, the body of a strong man and the head of a duck. You get the jist. Woody discovers that these mash ups are actually friendly, and can help, even though they are an unnatural combination.
With this in mind, below is a quick little combination of Pentaho and Jaspersoft, where Jaspersoft Reports are called from within the Pentaho User Console. This is admittedly not exactly rocket science, but should spark some thinking of where these disparate platforms can go when they are used collaboratively.
The basic steps to make something like this happen are:
- Make sure you have a working Pentaho server and a Jasperserver installed.
- Then, deploy the jasperserver webapp under the pentaho servers webapps path to allow the jasperserver to ride along with Pentahos Tomcat server.
- Next, within a pentaho solution folder of your choice. Add a *.url file to the solution that calls the Jasperreport of your choice. For the URL, add the decorate and jasperserver login parameters like this:
- And there you go:
Yes, there are some limitations here. It’s passing the plain text user information, and there is no common user model, and sessions are independent. It also adds some administrative complexity because you now have two different flavors of reporting under the hood, including separate database connections, etc. There are certainly sexier levels of integration with the jar files/class libraries possible.
But, simple integration like this can be a quick way to add certain capabilities across platforms, and speaks to the overall coolness that is open source BI. Yes, you can take the head of a pterodactyl and put it on a dolls body and you may find that it actually is helpful, and maybe not so scary.
Ill let you know if I get any more twisted ideas from watching Toy Story 3 this weekend.
This press release was recently shared with me and I wanted to call some attention to it. For marketers doing events, it’s a really impressive display of BI technology that’s worth taking a look at:
Eshots is a Chicago based event marketing firm that I’ve had the pleasure of working with in the past. One of the things that’s really nice about their event intelligence offering is how they see this as really a true technology platform that becomes and integrated component of their overall event marketing offerings. I’ve seen how they look at technologies like their Andriod-based event capture devices, and I think this will be a real differentiator in the market for them.
We all know that standard advertising isn’t as critical as it used to be, and eshots definitely has the right concepts in place to help advertisers get in touch and interact with consumers in new, untraditional ways, and most importantly, measure what’s working. You can read more about their event intelligence platform here:
It’s been some time since I last posted, and a lot has gone on with OSBI since then. With the 2.0 and 3.0 releases of Pentaho, there were major improvements to the UI, and the dashboard framework in particular. Now, with the pending release of Pentaho 3.5, significant improvements in reporting are on the way.
Pentaho Reporting has always been good, offering the baseline capabilities of standard banded report designing. But with the new Report Designer 3.5 release, I expect Pentaho Reporting will leap forward and fill many of the missing gaps in functionality and usability.
While all this is going on, there’s a major, and unbelieveable welcome push on the documentation side. For reporting geeks (you know who you are), there’s a new book that’s definitely worth getting your hands on. Pentaho Reporting 3.5 For Java Developers, by Will Gorman, is one of the first books available for users of the Pentaho Platform and it offers a complete view of Pentaho reporting from designing reports all the way through deploying them into various Java and web applications.
I think anyone who’s ever jumped into OSBI would agree, it can be hard to find clear “getting started” instructions. The wiki’s are good, there’s solid forums, and a great community, but not always an easily identifiable “first step.” If you want to get started with the innards on how reporting works, this book can get you headed in the right direction.
It starts out with a nice overview of the Pentaho reporting libraries, and then walks through a few tutorials on embedding simple reports into java apps. There are a series of deep dives on the graphical report designer that provide a nice illustration of all of the various elements that can be added to a report. Some of the new items in 3.5 (sparklines!) are reviewed here.
The book covers how to pull your data from all kinds of data sources, including SQL, OLAP, Kettle (Data Integration), Pentaho Metadata as well as XPath and HQL. There’s a nice chapter on charting as well. Some of the more advanced design items are covered, including the new parameterization model, sub reports and the cross-tab feature. The walk-through on cross tabs is particularly helpful. I don’t believe that I could’ve figured it out on my own, given the 3.5 UI is still evolving as of RC2.
There are several other topics in the book, including how to write reports with the API and extending reporting with custom functions. Meaty stuff, for sure.
On caveat on the book. Remember the title: “Pentaho Reporting 3.5 for Java Developers.” While it sure is a great resource if you’re doing Java development, it doesn’t mean that the book is only useful for Java developers. I’m not a Java developer, and I still find it’s a great reference and a offers clear guidance on the various properties and settings in Pentaho Reporting (Oh….that’s what “sticky” does!). That being said, if you’re a Java developer and get a thrill every time you type “ant,” this book provides some nice examples and demonstrations for embedding Pentaho into an application.
Should you pick this up if you’re not a Java developer? Absolutely. But, my word of advice is that if you’re not a Java developer and are just trying to get a sense of Pentaho Reporting, don’t freak out about the wealth of Java integration items this book walks you through. If all you want to do is “simple” BI reporting within the Pentaho platform, you’ll find a chapter in there just for you as well.
I like what this book offers. It’s got a nice intro to Pentaho reporting for novices, and then jumps deep into some hard core technical stuff, which really demonstrates the power of the platform. This book offers a great reference for those with experience deploying business intelligence, but are looking for some how-to’s on some of the finer points on the Pentaho platform. Kudos to Mr. Gorman.
If you’ve been looking at Pentaho at all, one of the things that has been tricky in the past was the ability to create slick dashboards, without the complexity under JBoss Portal. Most people expect to be able to access single page, interactive dashboards in their BI solutions. And while Jaspersoft will certainly make a leap with Jaspersoft v3, the Pentaho community has stepped up and provided the Community Dashboard Framework wiki.pentaho.com/display/COM/Community+Dashboard+Framework.
As I understand, Ingo Klose and Pedro Alves were instrumental in developing this, and they are to be congratulated for the awesome work. I’ve worked with the CDF for a few weeks now, and it’s a pretty impressive and easy to use framework. Once you get the hang of it, you can bring up dashboards fairly quickly. Note that it’s not for the faint of heart; there’s still a fair amount of getting your hands dirty with HTML and the glorious Pentaho XML files, but it does provide some very nifty tools for getting started.
There’s a couple things that I’d love to see added, including creating a pick list with a key-value pair and then being able to access the key for selection while displaying the values. We have a bead on how that might be changed. And the other limitation is the built-in Pentaho limit of 100 values on a pick list. I understand the Pentaho code could be modified to support a greater limit if you want to get into that.
There’s a couple other items that we’ve figured out while messing around with it. I’ve done some work on this, and we’ve added comments to the wiki (pending Pedro and Ingo’s review) on how to accomplish a couple other items:
1. How to add a change event to the dashboard components – wiki.pentaho.com/display/COM/Components?focusedCommentId=8290991#comment-8290991
2. How to add a manual refresh button – wiki.pentaho.com/display/COM/Gallery?focusedCommentId=8291017#comment-8291017
There’s a a few other techniques I’ve started to use to handle select all from pick lists, sorting the dashboard links at the top of the page as well as some page formatting things, but these are certainly a matter of style and preference on how to make the dashboard sing.
Again, kudos to Ingo and Pedro for making this kind of thing happen. It’s a nice way to get some flash to users and fill a gap that was sorely lacking in the Pentaho suite.
With the rapidly escalating capabilities of open source business intelligence software, many organizations are taking a look and exploring the technologies to see if it might be the right next step for them. But when initiating an open source business intelligence project, there are a few differences that are worth noting when compared with how proprietary BI applications are selected. The very nature of open source changes the dynamics of evaluating and acquiring software, and it’s worth understanding the different approach.
When evaluating an open source business intelligence (OSBI) project, there are some differences from proprietary BI applications selections that are worth noting. As with proprietary offerings, make sure you understand your priorities. Different OSBI applications are better at one area or another, so be certain you’ve captured what’s most important to you, and what’s next.
Be sure to do your homework, and understand what the products and the suite have to offer. Since these applications are modern by design, they are very lean, purpose-built from the ground up. On the flip side, though, also recognize they are often the composition of separate projects merged together under a common platform.
For an evaluation, you can explore the freely available software yourself. Open source software is ready for you to download and install, and you can see for yourself where it works for you. Don’t expect open source providers to participate in drawn out demos or to wine and dine you. In the open source world, you control your evaluation. And don’t be put off by some of the up-front complexity here. Most of these applications are very powerful under the covers, and some of that is not shielded from the user. It requires some technical know-how to understand the applications, but certainly, the power is there.
To prove it out, start with a pilot. As with proprietary applications, this is a good rule of thumb. But the real difference in the open source world is that you don’t have the barrier of acquiring the latest software to initiate a pilot. Since the software is typically available under General Public Licenses, the barriers are removed.
And as you complete a successful pilot, get the training, and get a subscription from the provider. There are many benefits that this education and support relationship will provide your organization in the long term, and it’s worth doing.
But don’t forget through all of this, the basics still apply. Good project leadership, engagement from both business and technical audiences, a solid staff, and sound data modeling and design will pay big dividends in your BI initiative.
I was talking with a colleague of mine the other day in the marketing database services business, and he was noting that there was an increasing interest in column-oriented databases for rapid database querying. This technology has been around for many years, and marketing database providers have used it for a while. Experian had a technology they acquired called Analytix that, for a time, was the dominant rapid query engine for marketing use. It blew the doors off of traditional technologies and was generally loved by users. It delivered quick counts from fairly complex questions.
As with anything, innovation continues, and other solutions for the same problem popped up. Bitmap indexes in commercial databases became the norm. Other concepts, like clustered databases, are solving the same problem. Greenplum is one of these very viable clustered databases, based on open source Postgres. Additionally, as noted earlier, new column-oriented databases with a more open standard are gaining attention. There are commercial players like Alterian, but also a new class of vendors, like Infobright. Also, look at Vertica, who is commercializing open source projects like C-Store. And there are other open source column databases like Lucid and Monet that are getting attention.
There are many approaches here to solve performance constraints in “non-traditional” database ways. And while many of these have lightning fast performance in getting data out, the biggest pain is in getting data in. These can all be very slow when loading data. In many cases, they rely on bulk loaders that handle data in massive chunks, rather than individual rows. I tried one column-oriented database on my laptop, with a load speed (granted, out of the box, no tuning) of something in the neighborhood of a dismal 2 rows a second, versus hundreds or thousands of rows a second for traditional databases.
Certainly, you can get the speed up by tuning these and applying the right kind of hardware to this, but if these don’t work well out of the box with common techniques, adoption of these kind of technologies will stifle. We’ve worked with some of these vendors and we can overcome many of the gaps, but it can certainly be a little tricky.
There’s a lot of promise in each of these areas. And with the right attention, I’m optimistic that the open source versions of these databases can be as successful as other open source database projects, like MySQL.
This is my first blog posting. So bear with me.
I thought the best way to start off would be to say “why do a blog?” There are a few reasons. First, it’s clear to me that the whole blogging thing is a great way to make your individual mark. Sure, I did tease bloggers at the time this whole “movement” started up, but it is a nice way to make public statements without hurdles and limits. This “internet” thing is pretty great, once you use it a bit. 🙂
When I did some work out on the west coast, the company I worked with really supported the idea of public communication with things like blogs. I like this concept. While I may agree that more information is not always better, the idea that organizations want their employees to share ideas in a public forum is great. It allows all kinds of people to contribute to the dialog, no matter who they are. And you never know where the next great idea may come from.
Also, I’m using this as a mechanism to improve my writing. As we increasingly work independently, face-to-face interactions are not as common as they are when worked together in the same location. Technology has enabled this, but we must remain vigilant not to let this isolate us. We need to take advantage of technology’s ability to allow us to communicate and share ideas, not just to allow independent thought. By better being able to communicate in this medium, I hope that I can contribute to the greater dialogue.
There’s just so many of us, each doing our own part. It’s nice to think that if we each recorded these thoughts and that we could access and share them, we could do something together that we never could have done separately.
And you never know. This might be the seed of the next great American novel.
Well that’s it for my first post. I’ll try not to wax so philosophical on my next posting. I did want to just lay some foundations out there so you know where I’m coming from, and where we can go together.