Trevor Wilford MBA

Have you seen this? I just tried it out this morning and it's pretty slick for spinning up a PBI report without leaving Jupyter Notebooks. I have a pro license and executed it in my personal Jupyter Notebook (not on Fabric). I have heard it does work on the trial licenses and in #Databricks. You can even push reports into workspaces for further use, collaboration, or presentation. Stay nerdy my friends! ------ Release notes: https://lnkd.in/eXUCAySG Documentation: https://lnkd.in/eU2peQqX ------ #powerbi #python #datavisualization

961

70 Comments

😱 I once made a really bad dashboard on purpose.   🏆 Then someone challenged me to explain how to fix it.   🔨 That was a lot of extra work, but I did it.   🖱️ I loaded it up with show/hide text buttons with explanations and links to external resources.   🚛 It's a truckload of data viz design knowledge in one place.   🔧 The dashboard has been adjusted after each of several presentations to university students in the U.S., Canada, and the U.K.   🏈 This is a Throwback Thursday viz!   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

12

3 Comments

How’s your journey toward AI panning out? More importantly the data inside your organization, is it ready for what the industry is asking of you? HUGE gap between ideation toward enabling while maintaining trust, confidence, governance, etc. The journey is real, and it’s happening. #AI #data #governance #confidence #management #security #compliance #privacy

10

1 Comment

❓ Do you know how big your text is on the screen?   🤔 I was trying to decide how much outer spacing to use in a Tableau text box, and I knew had 12 point font.   You can work for years in data visualization and not know how many pixels that is - because that was the case with me!.   It turns out that 1 point = 1.333 pixels, so my 12 point font is 16 pixels (or 1/3 bigger).   That made calculating my padding a lot easier.   👍 I hope that helps.   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

25

2 Comments

❓ Do you use the X/Y controls for small adjustments of your floating objects?   Or the arrow keys?   I am thankful that Tableau added the arrow keys for positioning.   😲 Because I tend to click the up control when I want the image to move up.   ▲ The up control *increases* the Y value.   🤔 And that makes the object move down (▼) on the dashboard.   How often do I do that?   It's my default behavior whenever I use the control. Yes, I'm neurodivergent, so I guess I'm just not like the usability test group.   So, I stick to the arrow keys for floating objects.   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

16

14 Comments

Check out my newest blog post on how to color the ends of a bar chart in #Tableau. One of the 5 simple design hacks I shared during a recent TUG talk. https://lnkd.in/eS7EJAxK

91

5 Comments

❓ Have you ever wanted nulls in Tableau histogram bins on the right?   The default is on the left.   In my use case, that didn't make sense.   I am showing the count of days of delay between two events.   So time increases to the right.   To me, it makes sense for the events that never completed (null) to appear on the far right - beyond the longest delay for something that actually happened, not before the earliest events.   One way to do this is to: ✔️Create a calculated field converting that null to a value greater than your biggest bin. Like: IFNULL([Days Delayed],99) ✔️Make that a discrete dimension ✔️Put it to the left of your bin pill so it sorts first ✔️Un-show headers ✔️Clean up formatting (like alias the null to "Never")   You can see the progression in the workbook link in the comments:   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

15

5 Comments

✔️ So you loaded your data into your viz tool.   🤔 Now, how do you start your analysis?   🪜 The 1st (and often skipped) step is to verify that the data arrived unscathed.   🧩 You may have had lots of joins across disparate sources that caused an issue.   🔧 Or maybe someone implemented a "fix" and didn't inform you.   Basic checks include: ✔️Number of Records of entire data source ✔️Universe of values within subsets of major dimensions (i.e. group by).    Same check within sub-dimensions ✔️Number of Records within subsets of major dimensions against known values in another tool (Excel, Python, R, SQL in database connection)    Same check within sub-dimensions (possibly as a crosstab) ✔️Validate at least a couple measures within dimensions - sums, means, ratios ✔️Trends include entire expected time range ✔️General scan for unexpected nulls   You don't have to do all of these every time, but simply forging ahead without any validation of whether the data arrived properly is guaranteed to bite you.   ⌛ It's just a matter of time.   ⛵ Happy Exploring!   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

3

3 Comments

❓ Have you ever wanted to include a SINGLE VALUE from another table with SQL?   👀 I saw some code from a coworker that was doing just that.   SELECT  * FROM table_a, (SELECT MAX (date_field) as max_other_dt FROM table_b) WHERE [some conditions]   This comma is ANSI-89 syntax, and it is forcing a cartesian product.   It's basically the same as: SELECT * FROM A CROSSJOIN B   (You could also do it as a LEFT JOIN or INNER JOIN on a condition that is always true like ON 1=1. There are probably additionalvariations.)   In our case, all the rows of the first table are multiplied by all the rows of the second table.   But the second table is not simply a single row; because we are using a single aggregate function, it's also a single column.   📑 The result essentially becomes a repeating final column on our first table.   If you run into this use case, now you know!   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

13

2 Comments

"Data is a precious thing and will last longer than the systems themselves." — Tim Berners-Lee In the data and analytics space, we often focus on the latest tools and technologies. However, the true value lies in the data itself and the insights we can extract from it. As professionals in this field, we must bridge the gap between technical innovation and business impact. Let’s continue to push the boundaries of what’s possible, but never lose sight of the business problems we aim to solve. Our work should always translate into actionable insights and tangible results. To business leaders and other professional across organisations, embrace the power of data-driven decision-making. Invest in understanding the potential of analytics and data science to transform your strategies and operations. Collaboration between technical and business teams is crucial for harnessing the full potential of our data assets. Together, just like my team at Turner & Townsend and I, let’s build a future where data not only unlocks some progress but creates lasting value. How are you leveraging data to make a difference today? #DataScience #Analytics #Leadership #BusinessIntelligence #DataDriven

32

1 Comment

After having spent majority of my #datascience career working on #nlp techniques, you realise the world of data science and within that, NLP (Natural Language Processing) is simply so vast. 🤯 Named Entity Recognition #ner was one such unexplored territory for me, and refreshing to learn how the predictions differ for this area of #nlp vs classic topic or sentiments detection. Working on this project with Marie Stephen Leo, sharing some of my humble learnings: 1. NER is classification done at granular “token”-level 🔤 within every sentence (document) 2. Predictions are “tags” (labels) within each document, such as “person”, “email”, “city”, etc. ☎️ 3. “person” tags can be further segmented into “B-per”, “I-per” tags 👩‍🦱 4. Tagging scheme can be “IBO1” or “IBO2” (IOB = Inside-Outside-Beginning). In IBO2, the B-tag is used at the beginning of every entity, regardless of context 🧑‍🦱 5. #zeroshot NER models such as GliNER, NuNER produce amazing out-of-the-box performance 🗃️ 6. There are cool evaluation libraries like #seqeval that can be used to evaluate performance of NER models and produce outputs like “classification report” with precision, recall, f1-score, similar to the good old sklearn ways 🧪 #ner #nlp #datascience #continuouslearning

60

1 Comment

What happens in the niche space where data ethics, data culture, and data literacy converge...with a sprinkle of AI hype? Stay tuned to find out.

6

⚖️ Have you ever wanted to have mark sizes be consistent across worksheets.   ❌ But you're not snapping to those convenient 50% lines.   ❌ And you lack privileges to edit on server?   🔨 Well, this was covered briefly in Tableau Conference, and I made an example dashboard - showing sizing with a parameter.   🤔 But you say, "Hmm. I want consistent sizing but I want it by a measure across worksheets."   I have an example for that too. Link in comments.   Of course, you typically wouldn't display this parameter for the user to control, but who knows? There may be a use case for that!   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

10

3 Comments

Try NLP Questions chat with census AI

2

❓ Can you put a dotted border on a bar in Tableau?   ❌ Not really, but you can make it look like you did.   A fellow Tableau developer at work had a requirement for a dotted bar on a Total of a Gantt chart, very similar to the image in this post.   You can't currently do that in Tableau.   But you CAN hack a look-alike: ✔ Create a separate worksheet with the total. ✔ Add it to you dashboard as a floating worksheet. ✔ Format the color of the worksheet to be the same as the bar. ✔ Add your border to the worksheet itself. ✔ Adjust the size (and continue to adjust by testing it on Server 🤣).   With this method, the total value is still dynamic, and with a centered label, still viable as a single bar.   A major weakness of this (as a Gantt chart add-on) is that it only really works for waterfall values that end positive and end in the maximum value - if the data is going to be refreshed. In those cases, you can size the total properly because Tableau will resize the elements to the same portion of the graph on the left every time.   The value that I sized to equal the net after the last two negative categories is unlikely to be the appropriate size on updated data.   However, it's fine if you're creating a final presentation.   Maybe this (or a variant) is useful to someone out there.   🏗→🧠 Build to Learn!   💭🚶‍♀️🚶‍♂️ Follow for more.   #tableau #data #analytics #VizoftheRay

15

7 Comments

🧐 Every data project comes with its set of challenges. Overcoming such hurdles is part of what makes working with data exciting and rewarding. My mom said I loved puzzles when I was little. The pieces look a bit different, but tables, joins, DAX, etc. are just parts of a larger puzzle. #DataQuality #DataChallenges

4

What do we get wrong as data leaders? Predictability. Let me explain. When your business partners ask you how long it is going to take to build a dashboard, what do you say? Commonly, you'll hear a range (like 4-6 weeks). And, also commonly, something throws a wrench in this range and it has to be extended. What about when it comes to investment decisions? For how example, how much will this new data platform cost? We think it will cost $1 million, and then after implementation it was actually $1.75 million. Businesses (and business partners) crave predictability. Now the detractors will say "data is too complicated for this" or "it's impossible to forecast exactly how much investment we'll need." Guess what? I don't buy it. If large technology companies can forecast out their financials, and can set dates for product releases, why can't you? Predictability is a way that a data team will have greater credibility and more likely to be treated as another business team. The more that you can do this as a data leader, the greater success you will see. #data #dataculture

22

7 Comments

Here’s a new article providing a simple example why more than statistical significance is needed for actionable inference #statistics

9

2 Comments