Jedi Tricks - Marc not Luke

This list will need further formatting, you might get the idea from this as it stands.

1. Mark Type leverage

Map with simple dot, mark type of bar not on maps, gantt bar will work on Map, conversion of the gantt bars to appear on the map.

Create groups to single out single data points with Other, drop the group on the colour shelf .

Small space visual filter.

Bar's with no axis, just use size only rows. Tableau works it out. nice.Use as filters tab on the dashboard to effect cross sheet filters.

2. Constant Reference Line - avoid the edit on the reference line

Create a parameter, show parameter control. Add parameter to the sheet, no calculation, add the parameter to the level of detail shelf.

Edit reference line on axis, use parameter.Anything in the level of details can be used to create reference line.

Synchronise axis - WINDOW_AVG() can be used against the means calculated

3. hide first item, first item of what from edit table calculation

index() -1

You can also filter table calculations

4. When did the view get updated, or data get extracted? <Data Update Time>

Using the title, insert tag to show when the extract was first made.or updated. Best information shown.

5. Migrate dashboard built against excel to a database without re-building the dashboard.

Export new data connection to a data file, replace connection required. Rename data source to match the one exported.

'Select Saved Connections'

Replace data source with new data source.

6.

Terradata - Migrate to the dwh or not?

Asked the question of how do you decide whether you need to report from transactional versus the dwh as the datasource? guess that I could have predicted the answer 'it depends'. Digging a little deeper two further suggestions, if you intend to merge datasets across silo then it makes sense to migrate the transactional data to the dwh and report. Otherwise performance could also be a good motivation, guess that you wouldn't really know this until you've decided to try it, unless your server runs so hot in which case an upgrade it needed anyway. Not really the decisive answer that I was fishing for.

Other interesting things, Terra data have partnered with Oracle around a CDC product, called golden gate, it allows you to set refresh rates between transactional and dwh that could vary. Useful feature to handle datasets that need support for different reporting frequencies.

Terradata also have a nice feature to move frequently used data to faster disk built into the product.

Also I seen to be the only person at the conference without an ipad2.

Stephen Few - Empower the Knowledge Workers

Key note on final day of the conference, Stephen Few talking about range of topics around visual perception.

Stephen predicates that Data Analysts are going to be the new scare skill, communicating the insights. Extract the story from the data he think will become the new value add skill.

http://www.perceptualedge.com/blog/

"Show Me the Numbers" -

Data Sense Making Process - Step 1. Search, Step 2. Examine, understand the story in the data Step 3. Explain the story to others. (SEE Process)

"Above all else show the data",

1. Reduce the non-data ink
2. Enhance the data ink

Chris Stolte - incremental updates

Spoke to Chris, he confirm that incremental update is just for insert, they plan to look at delete and update merge later. Such a shame, testing time is wasted.

Can see that this would only be useful for data that was never updated, e.g. tweets

The Pet Goat

Oh dear, Cheyne strike again.

Chris Stolte - six dot one deep dive

Chris goes deeper into the six dot one release

Data Performance

One up 550 million rows of data using the 6.1 data engine on, DELL laptop

Heard that the extract partial refresh only appends new rows, does not merge updates and or remove deleted rows. Unclear who would be interested in this, unless each row was a transaction with some action type? hard to then merge this in Tableau. Assumes that rows are never updatable?

Incremental refresh allow you to merge or append from different source, hmm interesting.

Credential pass through for SQL Server added to Tableau server. Yippie, we can finally start to use Tableau server. Cheyne feedback help push this one through. Also added for Terradata.

Interesting purge and pin feature of the start page, makes wokbook management easier.

Quick view of data on the Add Table and connection edit, as well as the join clause dialogue. Feature throughout the user interface. Preview results in the pane. Looks very similar to Excel.

iPad

Touch optimised application. No extra work to publish to the ipad.

Quick filter and selections have been revised to make it touch optimised.

ipad Tableau looks very impressive, support for all normal actions on the ipad.

New water droplet effect on ipad touch.

Key Note - Christian Chabot

Keynote address today from CEO and Tableau co- founder Christian Chabot, announced 6.1 features :

• Memory improvements for data engine
• Locale settings
• ipad app for native Tableau 

Seems 6.1 is a major release rather than a bug fix. Demo included some tick data from a European exchange, around 55 million rows. The data engine seems to now handle this in a breeze.

Locale settings handle dates and menus in range of languages, fairy standard stuff.

Demo of the new ipad app for native Tableau, fully compliant with tableau desk top. Initially for ipad but further support for more tablets in the pipeline. Careful attention to the way the UI renders for tablet use, particularly the way that quick filters are built for users.

Holy Cow

Wondering how to rotate objects in Tableau, delegate pointed me to this really cool dashboard. Sadly Tableau public spoils the party a little bit.

https://www.tableausoftware.com/fr-fr/about/blog/2011/01/3d-dashboard-its-not-what-you-think

Tool Tips - The icing on the Cake

Looking at the schedule, Andy Cotgreave from Oxford University is planning on a 1 hour talk on Tableau Tooltips. Wondering if there is really that much to say on Tooltips?

work shops

Matterhorn, Lausaane, St.Gallen? wondering what the connection might be, mountains or lakes perhaps?

Tableau have split the conference into 5 different categories of talk, core tableau, advanced applications, customer success, tableau in practice and hands-on training. I am trying to decide on what basis to select talks from each of these categories. Initially I think that an even split of talks from each category. However I am leaning more towards the practical side of the conference. Given the nature of Tableau product, rolling up the shirt and getting stuck in could be the best approach.

The Science of Visual Analysis by Jock Mackinlay looks like a good introduction to attend, get some of the fundamental principles straight first before we go any further. Might put myself down for that one.

Registration Complete

Arrived in Amsterdam yesterday evening, quick and uneventful flight. Read an interesting book by Ian Steward professor of maths at Warwick University on the flight, picked up a few interesting things in the book.

First one was called 'Most Likely Digit' - goes like this, if you look at a list of numerical data, and count how often a given digit turns up in the first digit in each entry, which digit is the most likely?

A quick Tableau worksheet in the departure lounge at Terminal 5 gives us the answer, it seems it's number 1? but why?

Second interesting thing was a joke ;

A biologist, a statistician and a mathematician are sitting outside a cafe watching the world go by. A man and a woman enter a building across the road. Ten minutes later, they come out accompanied by a child.

'They've reproduced' says the biologist
'No,' says the statistician, 'its an observational error'. On average 2.5 went each way.
'No, no, no' says the mathematician. 'It's perfectly obvious. If someone now goes into the building the building will be empty.'

Really liked this one.

Through registration, booked my technical walk through at 3pm today to get feedback from Tableau folks on my presentation ahead of the scheduled session on Wednesday.

Heading to attend the designing, building and maintaining dashboards workshop. Seems the conference is around 170 folks, more than a factor of ten smaller than the US version which is around 2000 this coming year.