A Data Science Central Community
Currently, common reports that we use are created in Crystal Reports Developer and then put into our ERP system. Users can specify some parameters and get an updated report as a PDF. Everything else is a request sent by email. A custom report is then generated and emailed back as an Excel file, which is laborious and slow.
I’d like to be able to offer users (product managers and the like) a simple way to perform some basic ad-hoc querying. I’m looking at Crystal Interactive Analysis, but if it’s like other products from BI/SAP I worry that it might have a steep learning curve. Also, at $495 per named user, it’s probably at the upper end of what I could justify. Datasource is MS SQL Server 2005, SaaS is out of the question, and I have a preference for webapps over desktop apps.
Have you heard of jasper reports / jasper server? It looks good to me, although I have not implemented it myself yet. The website can be confusing, but there's a full open source version. Commercial services are offered in addition to the open source version, but you don't necessarily need them. They have a lot of demo videos on the site. Sounds like exactly what you are looking for to me.
Don't sell the desktop apps short: as most users are still using spreadsheets having a deskside tool may be most productive.
Here are a couple of options you can take a look at that might expand your scope:
Google Refine (formerly Freebase GridWorks:) http://code.google.com/p/google-refine/
Tableau software: very slick if you ask me http://www.tableausoftware.com/
InetSoft: drag and drop dashboarding http://www.inetsoft.com/
I know you said no SaaS but just for other folks that are reading this thread and that are willing to embrace the future:
Some raw elbow grease solutions:
Pentaho Reporting http://sourceforge.net/projects/jfreereport/
JasperSoft Reporting https://www.jaspersoft.com/reporting
Now, I am still looking for a dashboarding solution that understands predictive analytics and can automatically adapt to the knowledge model that is extracted from underlying data. Tableau comes close but if you look at real information visualization it is still far more sophisticated as compared to dashboarding. Sort of the same difference between movie making and gaming: emotion plays an important part in communicating successfully and dash boarding with canned models is emotionless whereas leading edge information visualization is a like a good movie.
Thanks for the suggestions, I very much appreciate them. I found one more option that I’m considering – i-net Software’s i-net Crystal Clear (www.inetsoftware.de). Going to have to take a hard stare at all of them.
A couple of random notes… I’m not trying to sell the desktop apps short, they’re just more work to install, config, maintain, etc. I really think webapps are the way things will be done as technology moves forward.
As for SaaS, my issue is that I’m in a regulated environment and taking data out of that environment is difficult to impossible. As a general thought, honestly, I don’t know if SaaS is the future or not. Maybe I’m a bit fuddy-duddy on the issue. Potentially there are huge efficiencies to be gained from its use. I’d venture a guess that the average CPU load in our datacenter is maybe 20% during the workday and 5% if you took a 24 hour average. So many wasted resources and so much wasted electricity. But SaaS has huge hurdles to widespread use. Vendor lock-in is probably the most frequently considered. I think data privacy and security are much bigger issues however. If you look at companies that are purchasing SaaS, I have to wonder how many of them have a written contract signed by both parties, how many of them know who is responsible and liable in the event of a data breach, how many of them have any clue whatsoever what their vendors are doing to safeguard their private information. I’m sure most SaaS vendors do a good job with data security, but I’m also sure that there are some that are doing a really bad job. SaaS is a really scary proposition IMHO.
As for your dashboarding quest, good luck! Data presentation of a dashboard is straightforward; deciding what metrics to present and what boundaries those metrics should have to trigger dashboard changes is the hard part. Trying to do it an automated manner is a tall order.