📈 Data → Decisions✔ | Balancing Design and Pragmatism in Tableau | 2024 Tableau Social Ambassador
🧰 There are some standard practices you should have in your toolbox when working in Tableau. ✔️ For example, of you are going to perform major surgery on a sheet, duplicate it first. ✔️Likewise for calculated fields. 🔒 Do this because you want to keep a working copy around. Yes, you can back up the entire workbook, but you may have other changes in there that you want to keep. 🪙 Anyhow… 🔧 Another tool in the toolbox is brute force filter removal. 💾 For this, you want to save your workbook first because you are going to discard your changes after you're done with this one, and I'll explain why at the end. 🔎 Many times in analytics, your actual problem is one of filtering. If you suspect this to be the case, then read on. 📷 The image shows the number of filters I had on this sheet. Yes, it's a lot. 🌎 If you operate in the world of healthcare data and have convinced users to have 3 or fewer filters, by all means, enlighten me in the comments as to how to achieve that. 🧙 So, my Viz-in-Tooltip was returning no results (bonus tip: I put "VIT" in the worksheet name of all such sheets) . After A TON of fooling with filter parameters and online searches and videos, I finally channeled my inner virtual manager, Sandy. The real Sandy is quite amazing - she is both wise and powerful (to borrow an LOTR quote - but otherwise, she is unlike Saruman…), but she was out of the office at the time. 🔮 Inner Sandy told me to remove the filters one at a time until my data reappeared. (IRL Sandy had advised me to do this on a prior occasion). 🏆 This indeed revealed what the problem was. I was falsely assuming that two of the filters had no negative effects. (What did I post about assumptions and bugs?) 📝 The reason that I then just discarded changes is that because removing some filters (like those for All sheets using the data source) screws up all your other sheets. It's easier to just hack away, discard, and reopen. Yes, I *could* carefully remove and reinstate them, but that takes too much time and effort. 🏗→🧠 Build to Learn! 💭🚶♀️🚶♂️ Follow for more. #data #datascience #analytics #VizoftheRay
I love the filter removal method. I use it all the time when I get stuck.
I'm currently migrating some reports from QlikView (I swear, it's that's devil's own BI tool!) and reproducing dashboards with 28 filters!
Filters are often the culprit. Great tips here, Ray.
We used to have 10+ filters in reports. Now we have three and I havent experienced that it isnt enough. In reports we make three standard filters - based on parameter. Therefore, the user chooses the desired dimension in the parameter - and can then filter on the filter itself. It works very well
Yes! Always duplicate before troubleshooting. I’ll add: if you’re working with Tableau Online and are troubleshooting a Tableau Prep or a Tableau Desktop file… always download a copy locally and troubleshoot on your desktop. Yes, even for a quick fix. That way when you upload you can log what changed corresponding to the Revision log. Super helpful when you need to revert to a previous version.
Sr. BI Developer at Cotiviti
7moGreat tips. Filters are usually the first thing I look at when troubleshooting, basically doing as you described. I once inherited a dashboard with ~50 filters. I changed it to 5 core that everyone used often and 5 custom, where they could choose from a list. They agreed to the changes and of course wanted even more options in the list. It was a flexible solution that worked for them.