dashboard & report development process

34
1 www.data-intel.co.uk Dashboard & Report Development Process

Upload: duena

Post on 04-Jan-2016

140 views

Category:

Documents


0 download

DESCRIPTION

Dashboard & Report Development Process. Dashboard Definition. The fundamental challenge of dashboard design is to display all the required information on a single screen: clearly and without distraction in a manner that can be quickly examined and understood. An Effective Dashboard is a: - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Dashboard &  Report Development Process

1www.data-intel.co.uk

Dashboard & Report Development Process

Page 2: Dashboard &  Report Development Process

2www.data-intel.co.uk

Dashboard Definition

An Effective Dashboard is a:

Visual Displayof

the Most Important Information needed to achieve one or more objectives

whichFits entirely on a single computer screen

so it can be Monitored at a glance

• The fundamental challenge of dashboard design is to display all the required information on a single screen:

– clearly and without distraction– in a manner that can be quickly examined and understood

Page 3: Dashboard &  Report Development Process

3www.data-intel.co.uk

“How can I display this information in the most meaningful, clear, and efficient way possible?”

Layout of a Dashboard

Page 4: Dashboard &  Report Development Process

4www.data-intel.co.uk

Best practices while creating dashboards and reports

Dos -• Group data into logical sections• Condense data in form of summaries and exceptions• Concise, clear, and often small display mechanisms (sparklines, bullet

graphs)

Don’ts - • Using poorly designed display mechanisms (Pie charts)• Cluttering it with useless decoration(3D graphs)• Misusing or overusing colour• Supplying inadequate context for the data

Page 5: Dashboard &  Report Development Process

5www.data-intel.co.uk

Dashboard layout

• Different sections of data can be positioned on a dashboard depending on its importance

Page 6: Dashboard &  Report Development Process

6www.data-intel.co.uk

Designing Dashboard and Report

• Display mediums that can be used in a dashboard or a report are:• Tables –

» To look up and compare individual values» Data needs to be precise» To show multiple units of measure» To show both details and summary at the same time

• Graphs – » To show patterns and trends» To show relation among multiple values

• Icons – » To show any alerts» To highlight urgent or important data

Page 7: Dashboard &  Report Development Process

7www.data-intel.co.uk

Tables

• Highlight appropriately• Align data for easy comparison

Page 8: Dashboard &  Report Development Process

8www.data-intel.co.uk

Graphs

Commonly used graphs:• Bar graphs• Line graphs• Deviation graphs• Bullet graphs• Pie charts• Combination of bar and line graph• Stacked bar graphs

Page 9: Dashboard &  Report Development Process

9www.data-intel.co.uk

Bar graphs

• Used for nominal comparison between categorical subdivisions• Used for ranking

• Tip: Bar graph should always start with a zero value otherwise it could mislead

Page 10: Dashboard &  Report Development Process

10www.data-intel.co.uk

Line graphs

• Used to display time series or a frequency distribution with focus on the shape of the data

• Tip: Always place time dimension on the horizontal axis.

Page 11: Dashboard &  Report Development Process

11www.data-intel.co.uk

Deviation graphs

• Used when a measure is compared to a reference measure and to show the difference between the two values

Page 12: Dashboard &  Report Development Process

12www.data-intel.co.uk

Bullet graphs

• Used to display a single quantitative measure compared to one or more related measures (for ex. a target) and optionally with a qualitative scale subdivided into ranges (for ex. good, satisfactory and bad)

• Tip: Can be used effectively in dashboards to display achievement of target with a qualitative scale of good, satisfactory and bad.

Page 13: Dashboard &  Report Development Process

13www.data-intel.co.uk

Sparklines

• Used to display a time series with focus on the shape of the data and no need for quantitative precision

• It is ideal for dashboards as a means to display an historical trend leading up to the current measure, providing useful context in very little space.

• A sparkline chart can also be used to effectively show historical measures compared to a target, norm, or threshold.

• Sparklines are richer than trend arrows.

Page 14: Dashboard &  Report Development Process

14www.data-intel.co.uk

Pie charts

• Pie charts are used to show part-to-whole relationship• Arguably, pie charts are not the best medium to show the part-to-whole

relationship as difficult for eye to judge relative size of segments• It can be easily and effectively replaced by a bar graph

Page 15: Dashboard &  Report Development Process

15www.data-intel.co.uk

Icons

• Alerts - used to flag information as important or in need of attention. You can include variations of color to indicate different degrees of importance or urgency.

• Up/Down indicators - used to indicate that a measure is greater than or less than another measure (for ex. a prior measure in time to indicate that it has gone up or down)

• Note: Do not overuse alerts or icons as this will defy the purpose of using them.

Page 16: Dashboard &  Report Development Process

16www.data-intel.co.uk

Fonts and Colours

• Choosing a crisp/professional font can have a high impact on the reports.• For example:

• We prefer to use Calibri font for the content and Segoe UI font for the headings.

• If you prefer to use the company’s standard font, that could be used across all the reports.

• Further, you might want to consider what colour schemes you want in the reports. For ex. A particular colour for a brand or a country.

A table using Calibri font A table using Arial font

Page 17: Dashboard &  Report Development Process

17www.data-intel.co.uk

Standardisation & Business Rules

• It is ideal to set the standardisation rules across all the reports before starting to build the report. For example:

– Order of the parameters in all the reports, – All the percentage figures in all the tables should have only one digit after the decimal

point (90.1%) – The colours of the products in the graphs should be consistent across all the reports

• You can also define a glossary of terms to be used in the all the reports. For example, Cash sales should be called as “Value Sales”, Units sales should be called as “Volume Sales” and so on.

• Also, it is helpful to have all the business rules defined and documented before building the reports. For example, Market share for a product should be always be based on Market XYZ, growth is always based on Previous Year unless specified.

Page 18: Dashboard &  Report Development Process

18www.data-intel.co.uk

Report Specification

• The report specification document, will try to capture the exact layout and content of the report which has to be build.

• An Excel workbook can be used to put the layout and provide detailed information about the content to be displayed in the report.

• Please look at the Template for Report specs document for further details.

• The aim of having the report specification document is to avoid having multiple iterations of report development and make sure the report developed after the first iteration is very close to the final report.

• The document can be produced by the client or DI can produce it during the workshop (designated for dashboards and reports specifications).

Page 19: Dashboard &  Report Development Process

19www.data-intel.co.uk

Dashboards and Reports need to be personalised to the role of the end user so they get accurate and relevant information to them

A dashboard for National Managers will have high level information which will allow them to get a quick overview of the overall business.

A dashboard for Brand Managers will more detailed information which will allow them to have an in-depth analysis of their brands.

Page 20: Dashboard &  Report Development Process

20www.data-intel.co.uk

DI’s invested in developing a full report development process to ensure that the end user is bought in from first publication

Each

stage o

f this p

rocess is d

ocu

men

ted

and

invo

lves man

y ind

epen

den

t steps

Page 21: Dashboard &  Report Development Process

21www.data-intel.co.uk

1. Unit Testing2. Hierarchy checking3. Data consistency4. Chart linkages5. Double counting6. Custom calculations7. Error handling8. Business rules check9. Accuracy10. Labelling11. Client platform tests

Report testing

Involves at least 11 stages on it’s own,before a client sees the report for feedback

Page 22: Dashboard &  Report Development Process

22www.data-intel.co.uk

Charts Gallery

Ideally, you would want to have very simple and basic charts in the report without much of the effects (like 3D charts). This avoids the user from getting distracted away from the actual data shown in the report.

Page 23: Dashboard &  Report Development Process

23www.data-intel.co.uk

Charts Gallery (2)

More samples can be viewed at http://www.dundas.com/Components/Gallery/Flash/Chart

However, it is also possible to have some visually enhanced charts. Some of the charts that can be created in the reports are as follows

Page 24: Dashboard &  Report Development Process

24www.data-intel.co.uk

Types of InCell charts

The various types of InCell Charts that can be used are –

Line chart Column chart Deviation chart

100% Stacked bar chartBullet chart

Page 25: Dashboard &  Report Development Process

25www.data-intel.co.uk

Putting it all together

• When it comes to putting all the components together to build the final dashboard, make sure –

– The prime real estate on the screen is used for the most important data.– Small, concise display media are used to support the display of a dense set of a data in a

small amount of space.– Some measures are presented both graphically and as text.– The dashboard is not cluttered with instructions and descriptions that will seldom be

needed.

• Some of the examples that DI think are poorly designed, are shown in next few slides, giving the reasons why it is not an effective dashboard.

Page 26: Dashboard &  Report Development Process

26www.data-intel.co.uk

Poor Dashboard (1)

1. Unnecessary use of a dark background

2. Use of gauges is not ideal as it takes up relatively big space to show just one number. It can be effectively shown using just a number and an alert

Page 27: Dashboard &  Report Development Process

27www.data-intel.co.uk

Poor Dashboard (2)

1. Very little information shown.

2. Use of gauges, to show three numbers and taking almost half of the page.

3. Lack of context.

Page 28: Dashboard &  Report Development Process

28www.data-intel.co.uk

Poor Dashboard (3)

1. Unnecessary use of face pictures in the most emphasised area on the screen (top-right)

2. No definition of what the colours in the report mean.

Page 29: Dashboard &  Report Development Process

29www.data-intel.co.uk

Poor Dashboard (4)

1. Types of charts, difficult to interpret at first glance.

2. No context of numbers on the dashboard.

3. Gradient fill colours in the bar graphs add meaningless visual interest.

Page 30: Dashboard &  Report Development Process

30www.data-intel.co.uk

Poor Dashboard (5)

1. Overuse of gauges to show a few numbers.

2. Overuse of green colour through out the dashboard and hence losing the significance of green alerts in the table.

Page 31: Dashboard &  Report Development Process

31www.data-intel.co.uk

Effective Dashboards

• Some of the dashboards that DI think are effective are demonstrated in the following slides.

• Each dashboard are build for the targeted user and makes the best use of the real estate available to show the relevant and accurate information.

Page 32: Dashboard &  Report Development Process

32www.data-intel.co.uk

Sample 1 - Regional Dashboard

A dashboard is a visual display of the most important information needed to achieve one or more objectives; consolidated and arranged on a single screen so the information can be monitored at a glance.

Bullet chart shows the performance of the selected region against a set benchmark. It also provides a quantitative scale based on the different colours (eg. bad, satisfactory, good)

Deviation chart quickly shows the performance variance compared to the National numbers

Page 33: Dashboard &  Report Development Process

33www.data-intel.co.uk

Sample 2 - Territory Dashboard

Microcharts show the quarterly trend for each measure within a single cell to set the context around the numerical data being displayed.

Combined line and column chart shows detailed Sales vs Target trend for 12 months

A map plotting the location of patients in a given territory

Page 34: Dashboard &  Report Development Process

34www.data-intel.co.uk

Sample 3 - Territory Report

Finally, a recommendation is made in the ‘performance indicator’ column – for example, ‘must grow account’ where account market share and change in market share is below the territory figure.

A ‘must grow account’ at the top of the list is a very clear opportunity for improvement.

For reference the national and territory measures are shown at the top, followed by a structured table of accounts.

For the sales team we’re looking at here, the strategy was to work with defined accounts as much as possible. The accounts are customised groupings of bricks that are built into the underlying cube. These are then used to report on.

Accounts are ordered by market size (potential), and for each account we show the product sales, market share and change in market share.

A microbar chart is used to show the relative potential of each account against the average.