Let us review day 1:
- Most of the conference goers where relatively new to Tableau, few that I spoke with had more than a year or two of experience. There are two reasons for this. Firstly, Tableau has grown rapidly in the last year or two. The other is the conference seems geared toward newbies. It also doesn’t hurt that Tableau gives out free passes to new enterprise clients.
- Many of the presentations to on day one focused on data presentation and little on how to get the data into Tableau.
- There was little explanation on what insight people received from using Tableau. Most dashboards created were around user self service.
The conference started at 9 instead of 8 today. I think this was to give marry makers some time to recuperate after last nights festivities. The breakfast was different today, but just as good.
Session #6: Keynote: Neil deGrasse Tyson –Science as a Way of Knowing
Oddly, I wasn’t familiar Mr. Tyson. I should turn in my geek card. As co-creator and main host of the new Cosmos series I should have known who he was. The original inspired me to get into the hard sciences as a kid.
His presentation focused data should be used to explain the universe and how humans need to be open to new ideas that can challenge how we see the world.
Source: Tableau blog
- The tree of life doesn’t have humans at the top. We are just one of thousands of species on its branches. Also we are more closely related to mushrooms than to plants.
- It is not a matter of if, but when we will be hit by a large asteroid.
- We don’t have to destroy the asteroid to save the planet, only use the gravitational pull of a space ship to alter its course.
Takeaway: It might not helped me use Tableau better, but it was sure entertaining. Oh and Pluto is not a planet.
Session #7: Making a Viz that Stops Traffic
My first real session of the day brought in three Tableau Zen Masters, Paul Banoub, Anya A’Hearn, and Dan Montgomery. 20 or so are chosen each year based on the knowledge of the software based on their contributions to the Tableau community, user feedback, and employee nominations.
Thanks for Patty Walsh for snapping these.
The three all had decent presentation styles focusing on data presentation over analysis. As a fan of Edward Tufte, I wasn’t a fan of what they has built, but well-crafted. I felt they were too over produced to be practical for users. However, they all had great color sense and use of language.
- Use your dashboards to tell a story
- Pair down your charts using only ones that will enhance user understanding of the data
- Match colors and fonts with company standards
- If putting a visual online, use the Blog size, 650×650 canvas
- Max number of colors should be 2 to 3
- Use conversational language to explain visualizations
- Add pictures where necessary
- Tableau does not embed fonts, so you will have to screen print specialized fonts and add them as a graphics.
- Learn by pulling data on topics you love and building visualizations.
Takeaway: The is a trade off with visualizations. Data heavy ones might be easier for analysts but management and casual users need simple visually pleasing visualizations. Try to make your works look cleaner and less like Tableau.
Session #8: Porn, Pokémon & Pop Culture: Using Data That Doesn’t Suck
Similar to the previous session, a few more Zen Masters and a Tableau employee took the stage. For whatever reason, I thought that Jewel Loree, Andy Kriebel, and Peter Gilks had a better presentation. I think it was because they spoke less about their visualizations and more about how to get data into Tableau.
- Cut and paste is king. There are few tools that help clean data.
- Cleaning data it the most time consuming part of creating visualizations (hear-hear.)
- Excel is still comes before Tableau. You still need to build the data sets somewhere.
- KISS is the way to go as user engagement is the key to success.
- Create your own data to hone your skills. Use an app like Moves to capture what you do or last.FM to capture your musical life.
Note: since my pictures did not come out, I found the visualizations referenced in the presentation from from the presenter’s respective web sites. Click on the visual to discover more.
- http://import.io – An app that cleans up data from API so you can import it into reporting systems.
- http://www.reddit.com/r/tableau – Reddit user group. Lurk here for help.
- http://jewelloree.com – Go if you are a Pokémon fan
Takeaway: Tableau only makes data pretty. It is still hard to get the data in a useful format.
Session #9: Keynote: A Conversation with Michael Lewis
I was looking forward to this keynote as I have read a number of his books. I like how he brings the data users to life without getting into the gory details of the analytics they employ. Personalities always come before numbers.
This session was a Q&A that started off stiffly, but as he and Tableau’s Kelly Wright warmed to one another the conversation improved.
- His stories focus on personalities over data.
- Most of his stories are about outside or little guys fighting consensus in the marketplace.
- He related better to people who use theirs instincts vs detached people like venture capitalists. He finds it hard to connect with quants and entrepreneurs. The are too detached.
- He likes using a minimalist writing style and does not use jargon.
- Pairs down is story and leaves part unsaid to to let the reader interpret.
- Thinks that the right-left divide in the country was due in part to the financial crisis.
- The financial sector is scared of change. Many people fear technological change will cause them to loose their jobs.
- Thinks there is a disconnect in capital markets. Most people working on Wall Street are free agents and don’t hold allegiances to their employer of customers. This is harmful.
- He gets writing ideas from talking to people in his network.
Books I have read w/Twitter reviews:
- Liar’s Poker – Meh, skip it
- Moneyball: The Art of Winning an Unfair Game – Good to great, especially if you like sports and jerks.
- The Big Short: Inside the Doomsday Machine – Felt it was his best work, but then again I love the story of how big bankers got screwed.
- Boomerang: Travels in the New Third World – Good, a bit on the angry side and the most leftist of his books.
- http://data-informed.com/michael-lewis-talks-data-storytelling-tableau-conference/ – Another take on the conversation
- http://www.tableausoftware.com/about/blog/2014/9/michael-lewis-keynote-33001 – Tableau blog post
Takeaway: Work to pair away your data until you can see the story you want to tell.
Session #10: Use Tableau like a Sith
Much of the time, I spent wondering if I would use anything they presented in real life.
- Tableau employees hate pie charts.
- Tableau Data Stores are in XML. If one knows how to decode the parameters in the file, they can change a file saved in 8.2 to open in earlier versions, create pivot tables with more than 16 columns of data, change Excel connections to Access connections, and muck around to point at Internet URLs to pull in pictures rather than embedding them in the file.
- There is a way to create exploding pie charts using dual axis charts.
- There is a way to create gauges, but it is so complex why would you?
- You can use a jitter function to show over-lapping data points.
- There was a lot on how to improve Tableau Server performance, but since I am not a server admin, I didn’t bother to capture it.
- github.com/darthflashypants – Examples are stored here
- http://lightsiderecovery.com/ – Hack Tableau like a Sith
- Agent Ransack – I program that allows you to search of files anywhere on a network, including on the Tableau Server. The program is a personal favorite.
Takeaway: Tableau uses XML, so it is possible to do some cool things if you know how to read it.
Session #11: BI Reboot at Coca-Cola: Driving Sales with Tableau
Chathura Manawadu was wonderful. This was the best presentation of the day and a great note to go out on. He created a dashboard called SalesPulse for the thousands of Coca-Cola distributors that interact with the local Walmart and mom and pop stores.
- Designed their metrics from the bottom up as they implemented Tableau.
- Package the dashboard weekly by division. 3,500 users pick the packaged Tableau workbook off a file share and can use it when offline.
- Use a consistent color schema for all dashboards they create. Schema are color-blind friendly.
- Track sales, volume, and R.E.D. score.
- In the back end, use Business Objects connected to Teradata or Microstrategy connected to MS SQL databases and Excel files.
- Tableau blog post with Mr. Manawadu
Takeaway: Companies can use reader to distribute dashboard to field users. It just takes some effort to make it happen.
At night we took Seattle’s light rail to see a baseball game. I continue to be impressed by the cities transit. Chicago’s pales in comparison. There is little graffiti, it is cheap and it seems to connect bus, monorail (which I don’t think many ride,) and their funky trolley system well.
Safeco Field and Century Link field, where the Seahawks play are right next to each other on the south side of the downtown. Personally, I like the building aesthetic of the Seahawks field, but Safeco’s upper decks view has an unparalleled view of the city. It is the best view from any park I have been to date.
Garlic fries is a local Mariners fan favorite. The first third was overpowering, the second third wasn’t too bad…the last part was more garlic than fries. Vampires beware.
In the upper deck, it was a cold clear and clam night. The stars were out but the Mariners lost to the Astros. I thought their pitching was decent, but their hitting line up was scattered and weak. Luck is the only way they will sneak into the wild card playoff spot.
Update: 10/21/2016: Sine spell corrections.