The Best Graph

  • Comments posted to this topic are about the item The Best Graph

  • Steve, this posting brings back a funny memory of my last DBA position with a company that was purchased by Bridgestone Firestone a few years before I retired.  It makes a good point regarding the presentation of data via reports and graphs.  I spent eleven years there as a DBA and SQL Developer for a large VB/SQL Server project supposed to collect and distribute data 24/7/365 for a world-wide network of dealers.

    Early in this project I came across a series of ten reports that had been done by a VB developer, each of which was backed by a stored procedure.  These reports provided both internal company people and external dealers with data on customer transactions and could be requested interactively by users at will via the application.

    While there were no graphs involved, the reports themselves were pretty nicely done and presented data quite well.

    There was just one small problem.

    All ten of the stored procedures contained the same SQL design error which invalidatd the results.

    This project was supported by a project management group, a team or 6-7 developers, 5 DBA's, and a QA/testing group.  I dutifullly corrected all of the stored procedures and issued the change request to the project managers,  but that was as far as it went.

    I've been retired 13 years now but still communicate with some old co-workers.  The project itself was retired several years ago without ever having the invalid reports corrected.

    Moral of the story:  The best graph is a correct one.  Reports and graphs are worthless if the data is not valid.



    Disaster Recovery = Backup ( Backup ( Your Backup ) )

  • Very true. You need good data, which is why I advocate for testing. It's easier to read the requirements and see what was supposed to happen, and then you can compare that to the query/proc/etc.

    I can't tell you how many times I've worked with a customer that starts putting code in a VCS and seeing a bunch of it and then realizing there's a bug in their query that has been reporting bad data for years.


  • Good visualizations of data are essential, in my opinion. I came to that realization in my previous job. Most of the people I dealt with at that job were not good at analyzing raw data thrown up on a web page or report. But once you produced a graph of some kind, it made all the difference.

    However, the mindset in my current position is very different. I still think the majority of people aren't good at analyzing or deriving conclusions from a sea of numbers. But for whatever reason, dating back further than I've been here, they just do not like visualizations or graphs, unless the internal customer demands it. It's weird, but that's the culture here.

    Kindest Regards, Rod Connect with me on LinkedIn.

  • You have a weird culture.

  • Rod, at my age I like to use both good graphics and good summarized data.   Suffering from my OCD (self-diagnosed) I almost daily have to check in on my retirement investments.  I use Fidelity to manage my funds, and their website is fairly well done.  Once their after-market updates are done, usually by 7:00 pm they show up-to-the-day graphics for daily, month-to-date, year-to-date, life-of-account and even customizable timeframes for  results.  Then you can dig into the detail with a text presentation of gain-loss for all accounts and for  individual accounts and then further into individual trades.

    I won't get into the politics of this any further than to comment that over the last 2.5 years I've become hardened to seeing as much as four thousand dollars a day go down the drain.  This hurts regardless of bar-chart, line graph, whatever.

    Disaster Recovery = Backup ( Backup ( Your Backup ) )

  • Rick, I can imagine how a 4 grand drop can cause some anxiety. But it sounds like you've come to not be as anxious about it as it once was for you.

    Kindest Regards, Rod Connect with me on LinkedIn.

  • This was removed by the editor as SPAM

  • It's a control thing, Rod. If they have spreadsheets of numbers, they feel like they have the "raw data" in hand and can make decisions. If they have a graph (particularly if they can't sus out the answer or it's not a great chart for their question, like Steve's sample article linked) then they have to trust YOU, and they don't like that.

    It's a hard process to both educate them how to read different graph types, figure out what they're really looking for, and then convince them that they DO NOT NEED all of the data, they just need to see where things are going wrong. No point in paging through 49 states' multiple metrics just to find the one that went wrong. Create a graph of outliers, period.

    Please follow Best Practices For Posting On Forums to receive quicker and higher quality responses

  • Very good point, Jonathan. I think that is especially true for some of our users. A small, but significantly large number, of people are wizards at Excel. The reports we write tend to be columnar in appearance with tons of data. These power users will extract those reports into .CSV files, suck them into Excel and then slice and dice them the way they want to analyze the data. So, for those people they can produce their own graphs. Still, more than half of users just put up with lots of numbers on the page.

    Kindest Regards, Rod Connect with me on LinkedIn.

Viewing 10 posts - 1 through 9 (of 9 total)

You must be logged in to reply to this topic. Login to reply