Boulder Insight

Unlocking the Power of Data for Communities: A Call To Action

Deb Roth — Thu, 14 Sep 2023 16:13:48 +0000

Awareness, transparency, and engagement are hallmarks of a thriving community. Information – data – is what drives engagement. Traditionally, data about community issues and resources has been difficult to amass and present in an easy-to-read way. Thoughtfully-told data stories allow for more understanding, amplifying community awareness.

At Boulder Insight Group (BIG), we’ve long believed in the power of data to illuminate and drive meaningful change for individuals, organizations, and communities. We strive to create data visualizations that everyone can understand. That’s why we were delighted at the opportunity to work with The Colorado Health Foundation (CHF) for the fourth year to display their Pulse Poll data: a repository of navigable, in-depth information addressing core issues for Coloradans.

The era of obscure data and limited access is being replaced by an environment where clarity and truth are paramount. Presenting transparent and easily digestible data can invigorate communities, enabling informed advocates to rise. This dashboard beautifully demonstrates the ideal paradigm for engaged citizenry by ensuring data is transparent, public, and easy to grasp.

We invite you to explore the public 2023 Pulse Poll dashboard, as we believe it serves as a worthy template for all community organizations.

Takeaways from our CHF Collaboration:

Raising Awareness: Tableau-powered dashboards transform data into insights by making information engaging and easily interpretable.

Digestibility Over Volume: Rather than making anyone sift through a 50-page PowerPoint deck or hundreds of pages of spreadsheets, interactive dashboards can far more effectively make it simple to communicate, inform, and de-complicate large data sets.

Demographic Insights: The power of filtering results demographically offers granular insights into community sentiment.

The CHF Pulse Poll data reservoir is the power of informed decision-making that fosters robust community connections in acting. We think more community organizations should embark on a dynamic path: making data available and accessible to anyone who wants to see it in order to make a difference. Embracing transparency, fostering engagement, and empowering citizens and community organizations to pave the way for a brighter, more informed future.

Let us know what you think of Pulse in the comments below, and if you work for an organization or live in a community where you think this kind of resource could make a difference, comment below and we’ll reach out to start a discussion.

The post Unlocking the Power of Data for Communities: A Call To Action appeared first on Boulder Insight.

Don’t let Ventura turn you into a Tab-loo-hoo-zuh-her

Deb Roth — Mon, 28 Aug 2023 15:03:19 +0000

As a company, we ran into an interesting issue this week where our favorite datawarehouse, Snowflake, wasn’t connecting to Tableau Desktop. This was only happening with some of our employees – specifically those not using PCs.

After a lot of troubleshooting, we discovered that Macs currently running macOS Ventura couldn’t recognize the driver needed to access Snowflake. Obviously, that is a huge problem when considering data security.

Luckily, for those of you on MacOS, and specifically Ventura, we have found the fix to the Snowflake driver.

To be clear, this solution was not easy to find at all. It was a full day of research on Snowflake and Tableau forums with trial and error on a bunch of different options. The solution that worked for us (detailed below), is about #15 down on this link by someone named Wang Szeto: https://community.tableau.com/s/question/0D58b00009gWXTtCAO/snowflake-odbc-driver-doesnt-get-connected-to-tableau-desktop (there are few folks with that name on LinkedIn and we didn’t want to tag the wrong person but if you are out there Wang, please take a bow and our thanks for help with this!)

Crucially: You do NOT have to downgrade your MacOS and expose yourself to unnecessary vulnerabilities! For whatever reason, the Snowflake installer doesn’t make all of the necessary changes to your odbc files when you upgrade to the Ventura OS.

Here are 4 easy steps to manually make those changes and save yourself a lot of stress:

Download and install the universal package (not arm64) from here : https://sfc-repo.snowflakecomputing.com/odbc/mac64/index.html

Go to ~/Library/ODBC/odbcinst.ini and copy/paste the following text into this ini file:

[ODBC Drivers]

Snowflake = Installed

[Snowflake]

Driver = /opt/snowflake/snowflakeodbc/lib/universal/libSnowflake.dylib

Go to /Library/ODBC/odbcinst.ini and copy/paste the following text into this ini file:

[ODBC Drivers]

Snowflake = Installed

[Snowflake]

Driver = /opt/snowflake/snowflakeodbc/lib/universal/libSnowflake.dylib

Go to /Library/ODBC/odbc.ini and copy/paste the following text into this ini file:

[ODBC Data Sources]

Snowflake = Snowflake

[Snowflake]

Driver = /opt/snowflake/snowflakeodbc/lib/universal/libSnowflake.dylib

Once all 4 steps are done, when Tableau opens up and you click on Snowflake, it should point to the universal drivers.

Happy vizzing!

The post Don’t let Ventura turn you into a Tab-loo-hoo-zuh-her appeared first on Boulder Insight.

Friends Don’t Let Friends Struggle with Tableau: The One with Easy Naming Conventions

Deb Roth — Mon, 28 Aug 2023 14:58:34 +0000

I recently wrote about how some people often have the urge to verify data in Data Visualizations, and some steps we take to keep governance, and help people trust the data.

Despite that – people (finance…) – are still going to want to check. And for those folks, there are other things we can do in Tableau that speed up the process of checking complex calculations. We create QA sheets as a way to internally test our calculations that also allow the inheritor of the dashboard to validate however they see fit.

Tableau lets us create pretty cool nested calculations, even from levels of detail that are not in the view (AKA LOD Calculations). We have written some incredibly complicated formulas that always deliver the right answer but are hard for someone to dissect, whether in Tableau, or, god forbid, Excel.

The first thing to do when troubleshooting a complex calculation is to break it apart into smaller pieces and test each piece along the way. This process can be complicated and confusing as fields and calculations will often have very similar names. Tableau puts an = sign next to any calculation but I like to go one step further.

I group all calculations using numbers as a naming convention – that makes it far easier for me (and others) to read and process the calcs.

For example, instead of naming three calculations like the below:

SUM PROFIT
SUM SALES
PROFIT RATIO

I instead name them:

(_001) SUM PROFIT
(_002) SUM SALES
(_003) PROFIT RATIO (_001)(_002)

When looking at what field is used in a view, I can quickly reference the number of the calculation (001), versus having to read it to make sure it is the correct one, as typically there will be many calcs with similar names.

The other benefit to this naming convention is if I decide that I need to make a change to calc (_002), for example, I can type that into the search bar and not only will (_002) come up, but so will (_003) as the right hand side of the naming convention shows the calcs that are nested inside.

After 15+ years building Data Visualizations, I know that it is far easier and faster to find all calcs relevant to, say, SUM SALES when I look it up by a number (002) versus reading fields.

I’ve put the Boulder Insight Naming Convention Guide for Calculated Fields below in the comments (in the second comment) for anyone who might be interested.

Do you have any tips or trick for writing and finding complex calculations in Tableau quickly and easily? Please share in the comments – would love to hear.

GOALS

Simplicity of naming
Easy troubleshooting, especially when embedding calcs
Easy to teach/train staff
Easy to show/find total # of calcs in a given data source

NAMING FORMAT

(_00#) Name of Calc (_*Referenced Calc)
(_00#.P) Name of Parameter

PARENTHESES + UNDERSCORE

The Parentheses + Underscore is added when writing calc names into the calculation editor/dialogue box
Putting the Underscore on the search bar in the top left of your data pane will quickly show you only calculations, which allows you to see the total # of calcs in a datasource

NUMBER INDEX

The # is a unique identifier for each calc at the beginning – should be 3 digits (_001) (_021), OR (_201)

Start with _001 going up
If you delete a calc DO NOT fill it. Just keep going up with your numbering.
This helps you find calcs in pills on your sheets for easy updating and troubleshooting

NAMING

The names are descriptive of what the calc is doing.
They are not standardized like the numbers, so there’s no rules to try and remember

The post Friends Don’t Let Friends Struggle with Tableau: The One with Easy Naming Conventions appeared first on Boulder Insight.

Tableau and Chill? Only After a Double-Check with Excel!

Deb Roth — Mon, 28 Aug 2023 14:50:58 +0000

It used to be the norm for people to download data from Tableau then check it against formulas they kept in Excel. People are now more trusting of data visualizations (i.e. dashboards) but from time to time, they fall back to their old ways of checking the calcs, and they feel more comfortable doing that in Excel. (People in finance).

Rather than fight against that tendency we show our clients a few things that help them trust dashboards. We build out QA sheets, which comes from our preferred method of building calculations ‘in the sheets.’ This method breaks apart the calculation into the components of the overall calculation. That way, people who are so inclined can follow through on the math, the same way they would in an Excel File, but via the convenience of Tableau.

With the rapid advancements in AI and MachineLearning, data visualization will become even more powerful and immersive, offering unprecedented levels of customization and ever more-effective decisioning capabilities. It will seem easy than ever to engage with data, especially as we’re all getting more comfortable in using ML/AI tools to power daily decisioning.

That’s when our very human urge to verify data comes in. We shouldn’t discount that urge. We should validate it, and make doing so easy.

The post Tableau and Chill? Only After a Double-Check with Excel! appeared first on Boulder Insight.

I Can See Clearly Now

Deb Roth — Fri, 25 Aug 2023 19:18:42 +0000

We have several clients who insist on a light background with dark text. Despite my sunny personality, I always gravitate towards a darker background with light text, but that is not just because of look or feel. White is the brightest color, and our eyes are always drawn to the brightest color. People shouldn’t be looking at the dashboard background, they should be looking at what’s actually on the dashboard. That’s why I ultimately side with Dark Kermit: dark backgrounds allow for better focus on the data, and easier navigation, and that leads to more efficient decision making.

While there’s an argument that light backgrounds have their benefits, to me, the ONLY reason you would use a light background is if you were going to print your dashboard. (And if that’s your plan then I think having a dashboard might be overkill for your Lemonade Stand).

There’s a whole world of color options out there, though. Which Kermit do you side with when choosing dashboard backgrounds?

Article link of interest: Data Visualization Design: Dark vs Light modes in HappyFox BI

The post I Can See Clearly Now appeared first on Boulder Insight.

Bridging the Gap with a Guide

Deb Roth — Wed, 23 Aug 2023 18:25:24 +0000

When Training Data Just Isn’t Enough

Written by Heather Ditillo

I loved being a road warrior, traveling the country teaching Tableau and helping my students gain the confidence to be less intimidated by the technology and the skills they needed to better see and understand their data. When travel was curtailed in 2020 I started doing more consulting work with clients on their data projects in addition to teaching Tableau training courses virtually. And let me tell you, it was humbling to say the least, when our client’s real world data didn’t play by the same rules as the training data. I started to realize that training data is actually a fairy tale: clean, columnar and ready-to-go, but real world data is messy – really messy.

Learning Tableau using training data gets the job done for a certification program, but leveraging Tableau using real world data is a horse of a different color, as folks in The Wizard of Oz might say.

Often, when newly-trained students are tasked with their first data visualization project, they can have that “Toto, I don’t think we’re in Kansas anymore.” feeling (or sometimes flat out panic).

At Boulder Insight, we’ve discovered that students achieve greater understanding and proficiency in Tableau faster when we bridge the gap from training data to using their organization’s own data, messy as it may be.

Our most successful Tableau Training clients/students are the ones that have us to use their data in one of two ways (both are ideal):

1. By adding coaching sessions using their own projects and data to navigate the “ummm…this isn’t working like it did in training” situations;

2. Having us create a custom curriculum using a mock data set (with dimensions and measures like their company’s data). This further helps them gain practical proficiency and supports them on their path to certification.

We aspire to transform people into Tableau Superstars. As much as teaching, we focus on coaching: acting as a guide who helps people leverage their data to show the best strategic path forward, even after their classes end.

It’s as simple as this: training data has its place, but true mastery comes when you can apply those lessons to your own data.

Ultimately, people dive into data visualization initiatives to help their organizations make better decisions, faster. That’s the real value gained from using data they’re familiar with to get the expertise they need to let their data superpowers shine bright.

The post Bridging the Gap with a Guide appeared first on Boulder Insight.

Sugar-Coating Your Data

Deb Roth — Wed, 23 Aug 2023 18:23:24 +0000

When Radial Pie Charts Serve Up More Style Than Substance

Tableau Public is promoting radial pie charts and I don’t understand why. Radial charts are just tarted-up pie charts and pie charts are, on their own, terrible ways to communicate more than two pieces of data. Making a radial chart is like adding chaos on top of confusion. Has anyone ever seen, designed, or used a radial chart that facilitated understanding and action faster? (If so, please share in the comments! I’d love to see.)

Pie charts aren’t great for a variety of reasons but among the top are:

Determining actual figures is difficult
Labeling sections is difficult, and can end up looking sloppy
Too many pieces of data creates a visual unclear chart
You can’t use negative amounts in a pie chart

Radial charts can be visually beautiful, creative, and interesting, but as an effective visual communication tool they’re a crust, I mean bust. The whole point of data visualization is to communicate something clearly and efficiently, aesthetics are secondary. That’s why bar charts are so popular: they work. I’m not saying every dashboard needs to be only bar charts – there are tons of fun charts out there that can be visually more interesting than bars and still be meaningful, it just depends on what the goal of your data story is.

But when you optimize for style over substance, you lose exactly what you started off trying to achieve: understanding.

If you’re interested in reading more about this topic, check out this article by Steven Few.

The post Sugar-Coating Your Data appeared first on Boulder Insight.

Four Key Ways to Handle Tableaut (ta.BLOAT)

Deb Roth — Wed, 23 Aug 2023 18:16:21 +0000

tableaut /ta.bloat/

noun

def: When your Tableau Server has too many dashboards, making users too overwhelmed to make confident decisions.

“If we don’t consolidate these dashboards now, our case of Tableaut is just going to get worse.”

Similar: inDATAgestion, OverGraphed, PieEyed

Tableau makes it super easy to create content – that’s the good news. The bad news is that if you don’t revisit created content often, you end up creating slight variations of each (mostly for different audiences). This results in a condition I call, “Tableaut.”

Currently, there is no known ‘cure’ for Tableaut, however the symptoms can be managed with a few key efforts

Automate a weekly ‘archive stale content’ script that moves workbooks and data sources off the server and alerts the owners
Leverage the Logs and Postgres to determine what content is being used and by whom. Then organize a strategic consolidation / cleanup effort
Create Landing Pages where users can get to what they need from one spot instead of having to bookmark multiple links (here’s a good example of one we did: https://www.copulsepoll.org/results)
Run a post mortem on the most used workbooks, especially where you see users are downloading data. (If they are downloading data, then the dashboard is not giving them the answer they need.)

What are some other ways you handle Tableaut?

The post Four Key Ways to Handle Tableaut (ta.BLOAT) appeared first on Boulder Insight.

Knowledge-Driven Business Strategy

Deb Roth — Wed, 23 Aug 2023 18:00:39 +0000

You’re Going to Need A Smarter Boat

Data Driven Business (DDB) has been a buzz phrase for a hot minute. Although I agree in spirit of what DDB stands for, it just has never seemed quite right to me. I’ve been in the Business Intelligence industry for 20 years, and I’ve seen lots of clients chasing the elusive DDB grail. Time and time again executives realize they have too many dashboards that show their data, and that none of those dashboards are helping them make the best decisions.

The original intention of Data Visualization was born out of a need to analyze massive amounts of data in the quickest way. Leveraging the fact that the human mind can distinguish changes in patterns quicker using size, shape, and / or color rather than visually calculating the difference in numbers in a spreadsheet (my apologies to the finance folks who need their crosstabs). I am a HUGE fan of Data Visualization, but the fact is that Data has not and will not stop growing. Forget “You’re gonna need a bigger boat” – it’s now “You’re gonna need a SMARTER boat.”

While talking about how we help clients uncover actionable insights the team at Boulder Insight came up with a different term: Knowledge Driven Business. Because it’s not enough to be a Data Driven Business, what organizations need is the right knowledge, at the right time, to make the right decision, whether that’s a decision being made a few times a day or a thousand times a year. Clients don’t start out with that realization though: first and foremost, they think what they really want is dashboards that work well and look good.

What we’ve found most effective is realizing that while we’re all on a quest to soar, we have to crawl, walk, and run first. The path of our clients’ data journey (a journey ideally taken quarterly, or at least annually, because the only thing constant in business is change) goes like this:

Crawl | Business Strategy We ‘start with the why,’ by defining audiences, teasing out the right goals / KPIs, identifying decisions that need to be made and what information is needed to make those decisions.
Walk | Data Alignment We leverage what we learned in the Crawl phase to align, clean, and organize the ‘right data’ to answer the questions that will drive growth and increase value. Organization is critical as data needs to be BI-Readable (vs human-readable) and thus optimized for speed.
Run | Data Visualization and Training Once the data is organized from the Walk Phase, we either build or teach clients how to best build and use intuitive dashboards using Human Centered Design methodologies.
Soar | Automation The Automation phase is a BIG One! It’s where we all need to go. This phase is automating tasks through Scripting, Machine Learning, and Artificial Intelligence/Machine Learning, to hypercharge the human processing power and generate new datasets that provide the knowledge to make the right decisions at the right time.

Data will NEVER stop expanding, and we need all methods at our disposal to manage its growth.

Where are you on your Knowledge Driven Business journey?

The post Knowledge-Driven Business Strategy appeared first on Boulder Insight.

5 Ways to Improve Dashboard Engagement

admin — Fri, 13 Mar 2020 23:34:17 +0000

Sometimes it’s the little things that can make a big impact with data visualization. Odds are you are closer than you think to build amazing dashboards that will have a huge impact on your organization. We put together our top 5 ways to improve your Tableau dashboards that make an instant impact on your dashboards by improving user engagement.

1. Engage Your Audience

People are the key to unlocking your Tableau dashboard’s true potential and transform your business into a data-driven organization but a lot of the time people don’t seem to understand their needs or in some cases are actually afraid of data.

Start with your people. Take the time to learn about the audience and how they will be using the dashboards. Who are they? What’s their data competence? How do they currently report on information? What actions will they take on the data? What would make their job easier?

One great way to solve this is by building a dashboard the same way you would approach the product design of an App and apply IDEO’s Human-Centered Design Principles to the beginning of your project.

2. Simplify

Distractions add up. The human brain can only process 7 pieces of information at a time. This is the reason everything from a phone to credit card numbers is broken up into chunks of data so we can better remember and repeat this information. Cognitive overload is, unfortunately, a daily occurrence in our lives today.

Removing everything from your VIZs that are repetitive in nature or doesn’t tell your data story is key to achieve maximum clarity. A good rule of thumb is “if you said it twice you’ve said it wrong”. Work to reduce all clutter and duplicates in your dashboards.

3. Organize

When designing your dashboard layouts keep in mind that people need to focus on a starting point to explore your data story. UI and UX designers use a concept called a “hero image” to address users’ need to focus. It’s the main focus of the layout that everything else on the page supports. We can use this principle when designing dashboards as well. Build “Data Heros” by making the data view that tells the simplest and most important piece of your data story front and center by giving the most real state on the dashboard. All other data views should support the “Data Hero”.

Another great strategy is to design your dashboards like a website by placing navigation, filters, KPI’s and other key metrics in the same places that people are already use seeing. Keeping a consistent layout to all your dashboards will make it easier for people to engage with your data story.

4 Use Actions not Filters

The simplest way to gain performance and improve user engagement is to start using dashboard actions over filters. This not only reduces the data processing time but it enables and encourages people to directly interact with the data without reloading the dashboard or losing the relationship to all the other data points on the screen.

5. Data Story Colors

Colors have a huge impact on us, sometimes we don’t even know it happening. Take for Stoplight colors as an example, Red = Stop, Green = Go and Yellow = caution. Don’t fight it, embraces it! Use colors that invoke the feelings in people that people already have ingrained in their daily lives.

Also, take the time to design your data color story. Be sure to use the same color of the same metric every time. This way the user isn’t confused or has to relearn the data every time they see a new dashboard. And never use your corporate colors in the data story. Restrict these colors to your background template. You don’t want people associating success for failure metrics with your companies image.

The post 5 Ways to Improve Dashboard Engagement appeared first on Boulder Insight.