- What I mean by Data Visualisation
- Why Data Visualisation
- How it relates to software development
- Basic design guidelines and good practices about Data Visualisation
2. What we do
2
• Information dashboard for Business Intelligence
• Mission: Enabling better decisions
• Core values: innovative, open, supportive
About me
• Software Frontend Architect at Panintelligence
• Web Frontend Development for 7 - 8 years
• Responsible for reengineering Panintelligence dashboard
@panintelligence @mingzhi_huang
3. What we do
3
• Information dashboard for Business Intelligence
• Mission: Enabling better decisions
• Core values: innovative, open, supportive
About me
• Software Frontend Architect at Panintelligence
• Web Frontend Development for 7 - 8 years
• Responsible for reengineering Panintelligence dashboard
@panintelligence @mingzhi_huang
Twitter
9. 9
Introduction
Data
How do we learn?
Task estimation
Audit data
Timesheet data
Application data
…
@panintelligence @mingzhi_huang
10. 10
Introduction
Data
Are you making good
usage of your data?
Purpose
• show how to make use
of data visualization to
learn from data
• How to improve data
visualization
How do we learn?
Task estimation
Audit data
Timesheet data
Application data
…
@panintelligence @mingzhi_huang
12. 12
What & Why
@panintelligence @mingzhi_huang
How to improve
data
visualization
Colours Enhancements
What is data
visualization
Why data
visualization
Examples
13. 13
What is Data Visualization
Data
@panintelligence @mingzhi_huang
14. 14
What is Data Visualization
Data Data Visualization
@panintelligence @mingzhi_huang
15. 15
What & Why
@panintelligence @mingzhi_huang
How to improve
data
visualization
Colours Enhancements
What is data
visualization
Why data
visualization
Examples
16. 16
Why Data Visualization 1
@panintelligence @mingzhi_huang
To find out number of sales calls by Date
17. 17
Why Data Visualization 1
@panintelligence @mingzhi_huang
To find out number of sales calls by Date
Data
18. 18
Why Data Visualization 1
@panintelligence @mingzhi_huang
To find out number of sales calls by Date
Data Visualization
Data
19. 19
Why Data Visualization 1
@panintelligence @mingzhi_huang
To find out number of sales calls by Date
Data Visualization
Data
20. 20
Why Data Visualization 1
@panintelligence @mingzhi_huang
To find out number of sales calls by Date
Data Visualization
Data
21. 21
Why Data Visualization 1
@panintelligence @mingzhi_huang
To find out number of sales calls by Date
Data Visualization
Data
22. 22
Why Data Visualization 2
@panintelligence @mingzhi_huang
Slowly digest the information
No need to show you
street names at this level
23. 23
Why Data Visualization 2
@panintelligence @mingzhi_huang
Slowly digest the information
No need to show you
street names at this level
Shows street names
…
And some pubs
24. 24
Why Data Visualization 2
@panintelligence @mingzhi_huang
Slowly digest the information
No need to show you
street names at this level
Shows street names
…
And some pubs
Can you imagine
what the data
looks like for this?
25. 25
Why Data Visualization 2
@panintelligence @mingzhi_huang
Navigates information step by step to help digesting
26. 26
Why Data Visualization 2
@panintelligence @mingzhi_huang
Navigates information step by step to help digesting
Restrict to north
27. 27
Why Data Visualization 2
@panintelligence @mingzhi_huang
Navigates information step by step to help digesting
Restrict to north
Restrict to Newcastle
28. 28
Why Data Visualization 2
@panintelligence @mingzhi_huang
Navigates information step by step to help digesting
Restrict to north
Restrict to Newcastle
H Jones in detail
29. 29
Why Data Visualization 3
@panintelligence @mingzhi_huang
With a BI tool, the related SQL results can be grouped together
31. 31
Why Data Visualization - summary
@panintelligence @mingzhi_huang
Shows overall
pattern
Helps digesting
information step
by step
32. 32
Why Data Visualization - summary
Shows overall
pattern
@panintelligence @mingzhi_huang
Helps digesting
information step
by step
Groups related
information
33. 33
What & Why
@panintelligence @mingzhi_huang
How to improve
data
visualization
Colours Enhancements
What is data
visualization
Why data
visualization
Examples
34. 34
Why Data Visualization - Examples
Data Visualization only matters, when it shows
the information that is relevant to the users.
@panintelligence @mingzhi_huang
35. 35
Why Data Visualization - Examples
Data Visualization only matters, when it shows
the information that is relevant to the users.
@panintelligence @mingzhi_huang
Software Development
What do we care about?
37. 37
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
Sprint Burn down
Mike the Boss
20 days
sprint???
38. 38
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
Sprint Burn down
Points
completed by
developers
Tasks
completed by
developers
39. 39
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
Sprint Burn down
Points
completed by
developers
Tasks
completed by
developers
40. 40
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
Sprint Burn down
Points
completed by
developers
Tasks
completed by
developers
Boss cares
about
42. 42
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
I just like to work,
no need to sleep
???
K
M
Mike the Boss
43. 43
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
I just like to work,
no need to sleep
I just like to sleep, I
don’t want to work
???
K
M
Mike the Boss
44. 44
Why Data Visualization – Track Dev
@panintelligence @mingzhi_huang
M
Actually, 3 weeks
holidays at the beach
I just like to sleep, I
don’t want to work
45. 45
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
New requirements
???
Mike the Boss
46. 46
Why Data Visualization - Track Dev
@panintelligence @mingzhi_huang
New requirements
I don’t like
SQL
???
Mike the Boss
55. Colour for Exceptions
• Take a guess about what red means
55
@panintelligence @mingzhi_huang
56. Colour for Exceptions
• What would you guess orange means? (in the
same dashboard)
• Take a guess about what red means
56
@panintelligence @mingzhi_huang
64. Suggestions
• Better using the same colour to
represent the same meaning
64
Better using:
• Regular: muted
• Exceptions: bright
@panintelligence @mingzhi_huang
65. 65
How to improve
data
visualization
Colours Enhancements
Enhancements - what helps understanding
@panintelligence @mingzhi_huang
What is data
visualization
Why data
visualization
Examples
66. • Is 2014 performing well or not?
66
Enhancements - comparison
@panintelligence @mingzhi_huang
67. • Is 2014 performing well or not? • Compared to last year, what do you think?
67
Enhancements - comparison
@panintelligence @mingzhi_huang
68. • This shows the proportions
68
Enhancements - ordering
Performance of different cities
@panintelligence @mingzhi_huang
69. • This shows the proportions • How about providing
alphabetical ordering?
69
Enhancements - ordering
Performance of different cities
@panintelligence @mingzhi_huang
70. Performance of different cities
• This shows the proportions • How about providing
alphabetical ordering?
• How about providing
proportion ordering?
70
Enhancements - ordering
@panintelligence @mingzhi_huang
71. 71
Enhancements - scale
• What’s the difference between these two charts?
@panintelligence @mingzhi_huang
72. • What if we add the axes?
• Have you noticed the axes are actually the same?
72
Enhancements - scale
@panintelligence @mingzhi_huang
74. • How about these two?
• Have you noticed the starting points are different?
74
Enhancements - scale
@panintelligence @mingzhi_huang
75. 75
Enhancements - scale
When putting related information together
Would this be even more
confusing?
Option 1
@panintelligence @mingzhi_huang
76. When putting related information together
Would this be even more
confusing?
How about this?
76
Option 1
Option 2
Enhancements - scale
@panintelligence @mingzhi_huang
How does software development related to data visualization
If we want to consistently make good decisions, we need to have good evidence to prove what we do is right
We produce lots of data when we do software development. And the most reliable evidence is: data
Can you easily understand what’s going on? Probably not that obvious
Also, we need to scroll to see all the data
We wouldn’t be able to look at the information at once, how is it possible to squeeze all of them into a small visual space then?
Shows the pattern
Helps digesting the information
Shows the pattern
Helps digesting the information
Shows the pattern
Helps digesting the information
You now start to have new requirements: I want to know anything below 50
Instead of overwhelm you with all the data,
We can take steps to learn from our data
We allow people to navigate to the point that they care about the details
You don’t need to run SQL all the time, you can have the results side by side in their best visual representation
Value comparison
Proportion comparison
Change over time
Performance
I just want to show you the benefits, we are not here to talk about BI tools
Why data visualization:
Not just table – the result is the result of running SQL
Easier to digest
Shows summary vs. shows details – e.g. drill down
Shows details with table vs line
Results in new requirements
Shows patterns that tell a story you may not know – e.g. 2 lines comparison
Highlights problem – e.g. deviation chart
Why BI tools:
Remove repetition (grouping visual result of different queries)
No need to repeatedly write the same query
Consistently monitor the result of the same query
Show one table vs show multiple related charts
Why data visualization:
Not just table – the result is the result of running SQL
Easier to digest
Shows summary vs. shows details – e.g. drill down
Shows details with table vs line
Results in new requirements
Shows patterns that tell a story you may not know – e.g. 2 lines comparison
Highlights problem – e.g. deviation chart
Why BI tools:
Remove repetition (grouping visual result of different queries)
No need to repeatedly write the same query
Consistently monitor the result of the same query
Show one table vs show multiple related charts
Why data visualization:
Not just table – the result is the result of running SQL
Easier to digest
Shows summary vs. shows details – e.g. drill down
Shows details with table vs line
Results in new requirements
Shows patterns that tell a story you may not know – e.g. 2 lines comparison
Highlights problem – e.g. deviation chart
Why BI tools:
Remove repetition (grouping visual result of different queries)
No need to repeatedly write the same query
Consistently monitor the result of the same query
Show one table vs show multiple related charts
Result -> Why? -> is it because …?
Result -> Why? -> is it because …?
Data Visualization only matters, when it shows the information that is relevant to the users. So what do we care as developers?
- Scrum burn down - Estimate vs actual
- Tasks completed by developers
- Points completed by developers
- Estimate vs actual by tasks by developers
- Usage of features (although our customers told us it’s important)
20 days sprint? Or slow motion?
The Test People to help improving release process
Ken is the winner
Ming needs some kicking
Steve completes more points than Tiago
Tiago completes more tasks than Steve
= Steve worked on bigger tasks
Mike talk to K and M
- K said: I just like to work, no need to sleep
- M said: I just need to sleep, I don’t need to work (actually Ming is on holiday for 3 weeks)
While actually, M spent 3 weeks holidays at the beach looking at elephants
Audit data – stats of application usage
I’m about to show you some guidelines
Regardless if you use a data visualization tool – e.g. a dashboard, excel, Microsoft Dynamics
It’s always better to have these considerations
Why more important? - darker? Brighter?
Why more important? - darker? Brighter?
Why more important? - darker? Brighter?
e.g.
Red for negative / danger
Orange for warning
Green for good
Why more important?
What draws people’s attention?
What stands out more?
Let’s take a look at what sort of colour being used in these charts
You want the important information to stand out
You don’t want people’s attention to be on the less important information
The message:
Having comparison gives you a bit more context when trying to understand data
June is actually not that bad
The message:
Providing some ordering is better than not having anything