Ask HN:how to get BI reports from mysql
looking for some free tools/service, which are easy to use, to get dynamic reports out of mysql installation on ISP.
3 comments
For one of my customers, I have set up an ETL using ActiveWarehouse-ETL [1] in Ruby, which transforms the production data into a MySQL dimensional structure (facts joined with dimensions).
From there their non-technical users use plain Excel (their tool of choice) and an ODBC bridge to create reports as they wish.
See http://blog.logeek.fr/2008/1/19/a-beginner-s-guide-to-datawa... for interesting links I kept around, too.
[1] https://github.com/activewarehouse/activewarehouse-etl
From there their non-technical users use plain Excel (their tool of choice) and an ODBC bridge to create reports as they wish.
See http://blog.logeek.fr/2008/1/19/a-beginner-s-guide-to-datawa... for interesting links I kept around, too.
[1] https://github.com/activewarehouse/activewarehouse-etl
This is only a very basic dimensional model. It will not handle more sophisticated Business Intelligence reporting where serious comparitive KPIs are required.
You would have to hand build seriously comprehensive scripts to handle issues such as time dependency of attributes, hierarchies, navigational outliers.
If all that your business understands by BI are lists of data from the source or simple aggregation of current data, this is fine.
If they wish to analysis the historical data accounting for scenarios where the business process has changed over time this won't do.
You would have to hand build seriously comprehensive scripts to handle issues such as time dependency of attributes, hierarchies, navigational outliers.
If all that your business understands by BI are lists of data from the source or simple aggregation of current data, this is fine.
If they wish to analysis the historical data accounting for scenarios where the business process has changed over time this won't do.
Commenting a bit late, but: the model is indeed simple, but that's on purpose. It's actually the simplest possible model that fulfill the customer needs.
This system does indeed currently handle hierarchies, slowly-changing dimensions, and provides far more than lists of data, including some rather complicated KPIs.
In short: it fits the customer needs. Not to nitpick, but I've seen many people going with more complicated solutions (including SSIS and the rest of the MS suite) only to find out it was a major pita; that's why I'm advocating simplicity :)
Just curious: which kind of technical solutions have you been using and are happy with?
This system does indeed currently handle hierarchies, slowly-changing dimensions, and provides far more than lists of data, including some rather complicated KPIs.
In short: it fits the customer needs. Not to nitpick, but I've seen many people going with more complicated solutions (including SSIS and the rest of the MS suite) only to find out it was a major pita; that's why I'm advocating simplicity :)
Just curious: which kind of technical solutions have you been using and are happy with?
Happy with? Ho ho I'm never happy ;)
Look, I'm not going to say I love SAP BW - I loathe the costly bohemoth it has become, and I hate even more the so-called experts who are just out of business school with no CS knowledge, and those consultants who are "in it for the money". That stinks and there are many many of them.
But whenever I look at any competitive BI system, on a technical level SAP BW is way ahead of the game in so many areas. I can't even begin to list them properly:
Near code-ess datawarehouse
Follows the Kimball method to-the-letter
A truly usefull way of snow-flaking
and an astonishing delta load mechanism
source database independent scripting (if necessary)
Having said that only 1 client in the last 14 years has ever implemented SAP BW correctly and cleanly, and (incredibly) with no data load errors for 2 years. So I know how good it can be with the right person at the top (me;) it just costs the earth.
So to answer your question, the rest just look like very hard work, and almost all are by-and-large script instensive, which in the long run is unmaintainable because the business cannot understand it. And that is the whole purpose of a good warehouse - in my view - it needs to present a model that mirrors the business processes.
Look, I'm not going to say I love SAP BW - I loathe the costly bohemoth it has become, and I hate even more the so-called experts who are just out of business school with no CS knowledge, and those consultants who are "in it for the money". That stinks and there are many many of them.
But whenever I look at any competitive BI system, on a technical level SAP BW is way ahead of the game in so many areas. I can't even begin to list them properly:
Near code-ess datawarehouse
Follows the Kimball method to-the-letter
A truly usefull way of snow-flaking
and an astonishing delta load mechanism
source database independent scripting (if necessary)
Having said that only 1 client in the last 14 years has ever implemented SAP BW correctly and cleanly, and (incredibly) with no data load errors for 2 years. So I know how good it can be with the right person at the top (me;) it just costs the earth.
So to answer your question, the rest just look like very hard work, and almost all are by-and-large script instensive, which in the long run is unmaintainable because the business cannot understand it. And that is the whole purpose of a good warehouse - in my view - it needs to present a model that mirrors the business processes.