Following a discussion with Rody Zakovich regarding the constant requests from Clients to provide Cross-Tabs for all visualizations, we started to discuss the 'reasons' people want this, and what have been the most effective strategies to combat it...
Many clients have been embedded in Excel (or other similar text table) data analysis for years, and the cross-tab provides a 'security blanket'...I believe there is some sort of misplaced assumption (or lack of trust) where seeing the number in figures makes it feel 'more-real'/'more accurate', as well as a misunderstanding of the roles of 'Reporting' and 'Analytics' (which serve 2 different functions)....generally visualization is looking at trend, not perfectly audit-able exact numbers (i.e. you can't submit a bar-chart to your auditors). What people don't seem to grasp is any number (even an audit-able one) is always an approximation to something, which reminds me of one of my fave quotes
'An approximate answer to the right problem is worth a good deal more than an exact answer to an approximate problem.'
To try and get people to distinguish between 'audit-able accounts' and 'analytics' (during any Data Viz/Training I do) I always start with the same...
I get people to look at the attached document, and get them to look at it for 5-10 mins and ask them to tell me something interesting...usually get blank looks...then I show them this
btw the attached .csv is one I created as Hans has all the data he uses availiable on his gapminder site. I did try and recreate what he did in Tableau (using the page by)...however it soon transpired he did a lot more data cleansing for his TV show!! one day I'll get down to cleaning it all up a recreating fully in Tableau.
..and ask 'did not having the exact values in a table, inhibit your understanding of the data?', 'or 'did approximating the values as symbols [size, position, colour] help your brain to see the real pattern?'
...btw this is no guarantee...but I do what I can!!
GapMinder Combined.csv.zip 270.3 KB