James Q. O'Brien

I don't know about you #DataFam but I always look forward to Tableau Public's #VOTD email every day. I'm trying something new - I'm going to do a quick 60 second review of as many I can (it will not be everyday). Check out Zach Bowders, MBA's "Jay and Silent Bob Make Movies infographic build in Tableau. Nicely done Zach! https://lnkd.in/gYMXEBH7 #VOTDreview #GoForthAndViz

15

1 Comment

📣 I'm excited to be attending the upcoming Data & Analytics Live 2024, happening on July 17. 🎫 To join me, claim your free pass here: https://lnkd.in/e68KEunH #DataAnalyticsLive #data #analytics #Corinium #dataabdanalytics #webinar #onlinevent Corinium Global Intelligence

12

Look what arrived in the mail. Issue 4 of Nightingale, the journal of the Data Visualization Society. It’s all about emotion. ❤️‍🔥 Looking forward to a good read. #dataviz #datavisualization

16

4 Comments

Only two spots left for my upcoming dashboard design workshop! I’m Steve Wexler, founder, principal, and chief chart looker-atter at Data Revelations. I am one of three authors of The Big Book of Dashboards and the author of The Big Picture. I’m also a Tableau Iron Viz Champion and Tableau Visionary (Hall of Famer). I will be presenting my dashboard design workshop online on December 3 and 5. Learn more at https://buff.ly/3Bm0wnA PS: Don’t worry if you don’t use Tableau as the workshop is tool-agnostic. #dashboards #sigma #tableau #powerbi #excel

7

Big data is dead; long live reasonably sized data! I recently came across an insightful article on the rise of reasonably sized data that perfectly describes what I've been seeing for the past several years. Here are some of my highlights from the article. 🎈 Most people have an inflated estimate of their data volume: In reality, even a thousand customers placing hundred-line-item transactions daily result in data less than 1MB. A 16GB RAM laptop can handle three years' worth of this data. I recently created a 1 Million row, 100-column dataset, which was only 1.6 GB and easily handled by a laptop. 📈 Data storage requirements grow faster than Data processing requirements. Your data grows with new transactions daily, and you might have several years of data in your data warehouse. However, you're probably only querying and processing the last few months of data for reporting or ML models. Data storage costs are typically way cheaper than data processing costs, so you can cheaply add more storage as long as you have control over data processing. ✂️ Most queries process a significantly smaller volume of data. For instance, 80% of the data queried is likely from the last month and 99% from the previous year. Even with ten years' worth of data, you're probably only regularly querying 10% of it! Techniques like column projection, partitioning, and clustering can make data processing highly efficient by focusing on a smaller subset of the original data. The article claims that 90% of Big Query queries processed less than 100 MB of data, which is not exactly a "big" query! 🏋️ Hardware has improved tremendously. When the MapReduce paper was written in 2004, my personal desktop had 2GB RAM. Today, my laptop has 16GB! With efficient frameworks like DuckDB that can handle out-of-memory datasets, my laptop today can handle 75% of analytics tasks without the need for distributed computing like Spark. In conclusion, the rise of tools like DuckDB and Polars has put reasonably sized data at the forefront of data strategies. Newish roles like Analytics Engineers who can wrangle these reasonably sized datasets will play more prominent roles in data organizations in the near future. MotherDuck blog “Big Data is Dead”: https://lnkd.in/gWsjSTHZ I regularly post on data topics covering data science, ML engineering, and data engineering. Follow me to level up your data skills! Medium: https://lnkd.in/g2jAJn5 #analyticsengineering #dataanalytics #datascience #dataengineering #smalldata

96

14 Comments

I'm dreaming about overlay the chart with geographical map, so the histogram distribution will spread well along longitude-lattitude polygons

2

If you're in data science or product analytics, this is the knowledge you need to avoid deceptive metrics! My colleague and mentor Tessa Jones delivered an insightful talk on the Accuracy Illusion—where models seem to perform well but can mask underlying issues. She breaks down the pitfalls of relying solely on accuracy and reveals how we can dig deeper into performance metrics for better outcomes. #MachineLearning #DataScience #AI #ROI

6

Is taking snapshots of analytic results part of your data engineering toolkit? The impact of analysis and actions is often underestimated. Measuring systems have a hidden effect on results that may seem intangible until you see the progress made. #analytics #datawarehouse #measuredbenefit

2

Get an inside peek at Tableau Conference #Data24! Geekwire highlights the latest in AI and analytics, news from Tableau, interviews with the #DataFam, and more. Hear what's next for data and AI: tabsoft.co/4dL6xZF

7

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

The best data work is invisible. Until it isn't. When systems break Your documentation saves the day When questions arise Your process maps provide answers When decisions need data Your clean datasets enable insight When new team members join Your knowledge base guides them Your behind-the-scenes work Becomes front-and-center Exactly when it's needed most That's impact.

22

3 Comments

Data data everywhere...and oh, so advanced! Or so the headlines will have you believe. The reality is that many companies still operate with disparate systems for their data and data pipelines are barely in place. Each manager or report creator working from messy Excel sheets, often spending too much time getting things to work right, look right, or calculate right...only to have leadership ask for something different. 🤦‍♀️ Sound familiar? If so, this post is your chance to be a data hero for your company. 💪 3 Proven Ways to Excel ;) at Sourcing Insights 1. Template it Up Put together an easy template in your email, document, Slack, etc. that you can trigger anytime there's a request for a metric, data analysis, dashboard or whatever else. This template should ask things like due date, objective/"why" behind the request, and final audience. 2. Become Feature-Savvy Use Grouping, Data Validation, and XLookup to make your Excel life easier (and fun!). These allow proper organization and deviant values to be called out at a quick visual glance. Then, you can dig further into their why. This is important because the last thing anyone wants is irrelevant data mucking up important insights. 3. Structure Focus Focus on what's available, what could be attained, and what is not possible to know. Then, with what you have, mirror the likely questions to be asked in the order of columns. ❗ For example, will an account's customer lifetime value (CLV) be more important than the account's churn risk? Probably not. So, you can group all "churn" data together and place that before the account value metrics. Data organization and structure is actually the most important for analysis but hard to start with if you don't know your tools well. That's it, go forth and be a hero! P.S. Did these work for you or do you have others you'd add? We'd love to hear from you 👇 #ExcelSavvy #BusinessInsights #SmartProcesses #LeanThinking #Data

1

Great conversation today about data leadership! Big thanks to Celina Wong for thought-provoking questions. My biggest take-aways: 1) Data leaders in companies at every size and level of maturity have really similar issues - one of the biggest being the organization of data. Analytics Engineers/Data Architects are increasingly important to data success everywhere. 2) Leading with transparency and providing context and exposure are essential for retaining and developing talent - in any function, but in data especially. 3) Maintaining a balance between short-term priorities and longer-term infrastructure work is one of the most important skills for a data leader to develop

49

2 Comments

A fun #datavisualization that doesn’t get much attention. I think it visualizes distributions in an easier to read format than box-and-whisker plots alone.

4

As I'm building yet-another-presentation on how to structure an effective UTM parameter strategy, I remembered a blog post I wrote for @Tinuiti a couple years back on just this topic! Figured I'd share if you need a refresher.....or if you desperately need someone to use UTMs better and want to send them a not-so-subtle nudge. https://lnkd.in/gcXtE8Nb

30

5 Comments

AI, machine learning, and data, oh my! It's easy to get wrapped up in the excitement, but how do we think about the impact on humanity in parallel? #datagovernance #operations #secretops Check out the episode with Kate Parker! 🎥 Watch: https://buff.ly/3LurHy9 🎧 Listen: https://buff.ly/3JryjMz

3

My move into data management from analytics has been a great learning experience thus far, and I look forward to learning more. One aspect of data management is data quality, and I understand that data observability enables proactive data quality. Per Gartner's hype cycle, Data Observability is at the peak of inflated expectation. I see a clear value prop for data observability, but I'd like to know if any of you've succeeded in deploying data observability solutions with substantial proof of value. #datamanagement #dataobservability #dataquality #gartner

25

When I was a data leader, I broached the topic of data quality with the COO; they laughed in my face. In their defense, when there are a half dozen very legitimate business priorities, "data quality" does seem silly in a vacuum. This is not me taking a shot at that executive leader but a reflection I wanted to share with other data leaders because, frankly, as an entrepreneur, I get it, and data quality is not at the top of my list for my organization. If I were to guess, 9+ out of 10 executive leaders only care about data to get quality information quickly. Most do not see data as an organizational transformative topic. Is this their blind spot or an opportunity for data leaders? Rather than present data quality as a problem. Present how a business leader needs help to run their process more effectively by improving their reporting capabilities, which may improve data quality. It's a subtle change in message, but hopefully, you won't get laughed at! Hopefully helpful to someone!

17

5 Comments

Are you confused by all of the hype around data products? If yes, you should check out the latest episode of the CDO Matters Podcast! On this episode I demystify all of the hype, and provide clarity on the various perspectives around data products. Data products exist on a spectrum, from raw materials on one end, to consumable solutions that solve end user problems on the other. If you are considering data products, I also provide my perspective on the best way to go about implementing them. CDO Matters Episode #62 is available now! #dataproducts #cdomatters #cdo https://lnkd.in/edbRuch3

42

11 Comments