how do we design dashboard? are we suppose to integrate everything ? is there a need for web report?
What Is a SAS BI Dashboard?
A SAS BI Dashboard surfaces information in a graphical format. Dashboards are intended to provide information that can be evaluated quickly. In many cases, they enable you to view more detailed information by clicking on a dashboard item.
Building and Using Dashboards
Creating dashboards is typically handled by a power user. The SAS BI Dashboard application is used to create and maintain the SAS BI Dashboard components. Viewing and interacting with dashboards is something that is generally available to anyone with access to the SAS Information Delivery Portal.
What Is the SAS BI Dashboard Portlet?
The SAS Information Delivery Portal enables you to use the SAS BI Dashboard Portlet to view and interact with dashboards in the portal framework.
Dashboard Components Dashboards consist of several components. These components are objects stored in the metadata repository. Example of components are:
A dashboard is a container that displays one or more indicators or static objects. The dashboard can be displayed in the SAS Information Delivery Portal or in the SAS BI Dashboard application.
An indicator is an object that defines the display settings and other properties for visually displaying information in a dashboard.
The Indicator Data component retrieves data for an indicator. This component is basically a query against the data source
A range defines the measurement intervals by which a metric is evaluated. In addition to the code intervals, the range also defines strong text the colors and labels displayed in an indicator.
Before Building a Dashboard
Before you begin to build a dashboard using the dashboard designer, you should spend time designing the dashboard.
Determine what type of information the dashboard needs to convey.
Choose the indicators that you will display in the dashboard.
Locate or create the necessary data sources. The indicators that you choose determine the type of data that you need. You might need to create new data sources, calculated fields, or summary tables.
Steps to Build a Dashboard
After you design the dashboard, the following steps are typical for building the components to implement your design:
Define indicator data .
Create a range (if necessary).
Create one or more indicators .
Create the dashboard and define interactions.
Test the dashboard and its corresponding components
Reference: Blackboard SAS Practical 9.
Dashboards are often created on-the-fly with data being added simply because there is some white space not being used. Different people in the company ask for different data to be displayed and soon the dashboard becomes hard to read and full of meaningless non-related information. When this happens, the dashboard is no longer useful.
This article discusses the steps that need to be taken during the design phase in order to create a useful and actionable dashboard.
Think about the audience for the dashboard. The most effective dashboards target a single type of user and just display data specific to that ‘use case’.
Is the dashboard going to be used by the executive team to monitor the company financials or will it be used by the marketing team to monitor daily activities? It’s important to ensure that where possible your dashboard consists of data that’s specific to a single audience. Often this step is overlooked and dashboards include a mix of data: Some of which is relevant to one audience and some to another.
There are 3 common types of dashboard, each performing a specific purpose.
The types of dashboard are:
Operational Strategic / Executive Analytical
Operational Dashboards These dashboards display data that facilitate the operational side of a business. For example, in a business with a website, it’s important to ensure that your website remains up and running, so you would monitor server up-time and utilisation. In a business with an inside sales function, you may want to create a dedicated sales dashboard that displays number of calls made and number of appointments booked.
Think of an operational dashboard as monitoring the nerve centre of your operation. Operational dashboards often require real-time or near real-time data
Strategic / Executive Dashboards Strategic dashboards will typically provide the KPIs (Key Performance Indicators) that a companies executive team track on a periodic (daily, weekly or monthly basis). A strategic dashboard should provide the executive team with a high-level overview of the state of the business together with the opportunities the business faces.
This data could be: Periodic revenue (vs prior period) Costs (vs prior period) Headcount (by department) Sales pipeline
Analytical Dashboards An analytical dashboard could display operational or strategic data. However, this type of dashboard will offer drill-down functionality - allowing the user to explore more of the data and get different insights. Often dashboards include this functionality when it is not required. Do not simply provide this functionality because you can.
Bear in mind that different user groups may require a different type of dashboard. The marketing manager may need both a Strategic and Operational view of their data. Where possible create two separate dashboards.
A well-designed dashboard will ensure that data is displayed in logical groups. For example, if a dashboard includes Financial KPIs and Sales Pipeline, ensure that the financial data is displayed next to each other, with the Sales Pipeline data displayed together in a separate logical group.
Grouping is often by department or functional area and can include:
Product (Inventory, development)
Finance (Actuals and forecasts)
Often the most important real-estate on a dashboard (top left-hand corner) is reserved for a company logo or a navigation tool. This is not good dashboard practice as the part of the screen is the most important part of your dashboard (this is because most western languages read from top to bottom and from left to right - hence our eye will start it’s journey when discovering something new at the top left-hand corner.
An Executive dashboard can have a number of different audiences. Ensure that the data you display is relevant to the users. Think about the scope and reach of your data:
The whole company By Department Individuals Suppliers Ensure that you understand exactly who the intended audience is and the scope of their requirements. In a small organization, the Executive dashboard is likely to include KPI data across all departments. However, in a larger company, each department may have their own Executive dashboard.
Gaining agreement on dashboard components from non-related parties is doomed to fail.
Dashboards are often cluttered. Cluttered displays deflect the focus from the important messages. Some are cluttered with useful and relevant information and some are cluttered with useless and irrelevant information. Neither of these situations are desirable.
Each dashboard type may require different amounts of data (for example an Executive dashboard may only need 6 numbers, whereas an Operational dashboard may need upwards of 20) There is no hard and fast rule to follow here, except ensuring that everything you display is relevant and meaningful to the audience. Do not add a graph or text simply because you can.!
** Ensuring that your dashboard data is being refreshed at the right intervals saves time during development (why go through the pain of sourcing real-time data, when all you need is a weekly feed) and can ensure optimal performance once the dashboard is live.
Examples of refresh rates on dashboards include:
Real-time (or near real-time) Daily, weekly, monthly
As a rule of thumb, operational dashboards require data in real-time or near real-time, whereas executive/strategic dashboards require data refreshed on a less frequent basis.
Some more tips Provide context
How will you know whether those numbers are good or bad, or whether they are normal or unusual if there is no context? Without comparison values, numbers on a dashboard are meaningless for the users. And more importantly, they won’t know whether any action is required. Always try to provide maximum information, even if some of them seem obvious to you, your audience might find them perplexing. Name all the axes and add titles to all charts. Remember to provide comparison values. The rule of thumb here is to use comparisons that are most common, for example, comparison against a set target, against a preceding period or against a projected value.
Make it simple
Nowadays, we can play with a lot of options in the chart creation and it’s tempting to use them all at once. However, try to use those frills sparingly. Frames, backgrounds, effects, gridlines… yes, these options might be useful sometimes, but only when there is a reason for applying them. Moreover, be careful with your labels or legend and pay attention to the font, size and color. It shouldn’t hide your chart, but also be big enough to be readable. Don’t waste space on useless decorations, like for example a lot of pictures.
Be fun and creative
This point seems to stand in contradiction to what we have already said. However, when we stressed that the colors should be subdued and the layout well-thought-out, we didn’t mean that your dashboard should look boring. On the contrary, we want you to let go of Power Point style presentations from the 90s. The modern dashboard is minimalist and clean. Flat design is really trendy nowadays.
Don’t go over the top with real-time data
Don’t overuse real-time data. In some cases information displayed in too much detail can only be a distraction. Unless you’re tracking some live results, most dashboards don’t need to be constantly updated. They serve as a picture of a general situation or a trend. Most project dashboards must only be updated periodically – on a weekly, daily or hourly basis. After all, it is the right data that counts the most.
The context and device on which users will regularly accesses their dashboards will have direct consequences on the style in which the information is displayed. Will dashboard be viewed on-the-go, in silence at the office desk or will it be displayed as a presentation in front of a large audience? Remember to build responsive dashboards that will fit all types of screens, whether it’s a smartphone, a PC or tablet. If your dashboard will be displayed as a presentation or printed, make sure it can be contained within one page.
Each dashboard should be designed for a particular user group with the specific aim of assisting recipients in the business decision-making process. Information is valuable only when it is directly actionable. The receiving user must be able to employ the information in his own business strategies and goals. As a dashboard designer who uses only the best dashboard design principles, make sure you are able to iden
tify the key information, and separate it from the inessential one to enhance users’ productivity.
In conclusion , By following the simple steps explained above, your dashboard will be well designed and only contain relevant data that will generate the insights that you need.
A good dashboard is one that
Makes the complex simple: we have lots of information, lots of data that changes all the time and different analytical needs and questions. We want to take all this complexity and make it simple.
Tells a clear story: we want to be able to connect data to its context in the business, and to answer the viewer’s questions. This is where the visual layout of a dashboard plays a crucial role.
Expresses the meaning of the data: the chosen data visualizations need to correctly represent the data, and the information you want to extract from it.
Reveals details as needed: we want each viewer to have access to the data they need – no less but also no more. Some users might need to be able to see a more granular view of the data – others could suffice with an overview.
Here are some tips to creating a good dashboard .
Don't make my mistake! Clearly name and explain the metrics, units of measurements, and values shown in your visualization. And keep it consistent for all of your visualizations.
Red means bad, green means good is easy for people to get. When you start adding in other colours like blue for improving, orange for declining, yellow for neutral, and gray for no change, things get confusing. Keep your colour scheme simple.
The same goes for display icons. Stick to a handful of icons and keep it consistent across all of your visualizations. If you use a green check mark to signify a positive value in one visualization, don't use a green square in your next visualization.
Avoid using multiple types of time frames on a single data dashboard, such as last 30 days, this quarter, and last year. If you want to show historical data alongside a rolling time frame, make sure the time frame lines up and is clearly labelled.
Showing a date on the dashboard is a great way to provide context. Try and stick to a single format to make that information easy to process.
We made $32,435,546.12 last year! Instead, you made $32.4M. If an end-user wants more detailed information they can either go to the source, or view a dashboard that provides much more detailed information.
If the data doesn't add to your story, ask yourself why you're showing it. The exception is when your audience demands more data.
Can't avoid showing a complex data set in its entirety? You can still be kind to your end-users and use drop-down menus to allow them quickly sort and filter the data they are viewing.
Consider how the dashboard is going to be viewed. Is it going to be displayed on an LCD TV? Or, it is going to be viewed on a mobile device? The differences between devices can have a huge impact on dashboard consumption.
What does your audience want from the dashboard? Awareness of your audience, their data reporting needs, and how they will be using the dashboard should guide all of your design decisions. An executive will want a very different dashboard than a business analyst, and you need to design accordingly.
Your dashboard should provide the relevant information in about 5 seconds.
Your dashboard should be able to answer your most frequently asked business questions at a glance. This means that if you’re scanning for the information for minutes, this could indicate a problem with your dashboard’s visual layout.
When designing a dashboard, try to follow the 5 second rule – this is the amount of time you or the relevant stakeholder should need to find the information you’re looking for upon examining the dashboard. Of course, ad-hoc investigation will obviously take longer; but the most important metrics, the ones that are most frequently needed for the dashboard user during her workday, should immediately ‘pop’ from the screen.
When designing a dashboard it’s important to follow some kind of organizing principle. One of the most useful ones is the inverted pyramid (see image). This concept originated from the world of journalism, and basically divides the contents of a news report into three, in order of diminishing significance: the most important and substantial information is at the top, followed by the significant details that help you understand the overview above them; and at the bottom you have general and background information, which will contain much more detail and allow the reader or viewer to dive deeper (think of the headline, subheading and body of a news story).
How does a journalistic technique relate to dashboard design? Well, business intelligence dashboards, like news items, are all about telling a story. The story your dashboard tells should follow the same internal logic: keep the most significant and high-level insights at the top, the trends, which give context to these insights, underneath them, and the higher-granularity details that you can then drill into and explore further – at the bottom.
Choose a few colors and stick to them . The important thing is to stay consistent and not use too many different colors. You can choose 2-3 colors, and then play with gradients. A common mistake is using highly saturated colors too frequently. Intense colors can instantly draw users’ attention to a certain piece of data, but if a dashboard contains only highly saturated colors, users may feel overwhelmed and lost – they wouldn’t know what to look at first. It’s always better to tone most colors down. Dashboard design best practices always stress the consistency when it comes to the choice of colors. Use the same color for the same item on all charts. It reduces the mental effort on the users’ side and makes dashboards more comprehensible. Moreover, if you want to display items in a sequence or a group, don’t just go for random colors. If there is a relationship between categories (e.g. lead progression, grade levels etc.), you can make it easier for users by using the same color for all items, but graduating the saturation. Users then only have to remember that higher-intensity colors symbolize that the variable displays more of a certain quality, which is easier than memorizing multiple random colors. Again, what we are trying to achieve is creating a dashboard that can be understood by the user as quickly as possible. Our last suggestion when it comes to colors is to be careful with the use of “traffic light” colors. For most people, red means “stop” or “bad” and green means “good” or “go.” This can be very useful when designing dashboards, but only when you use these colors accordingly.
Make it as easy as possible
Don’t lose sight of the purpose of designing a dashboard. You do it, because you want to present data in a clear and approachable way that facilitates the decision-making process. If you make the charts look too complex, the users will spend even more time on data analysis than they would without the dashboard. Data analysis displayed on a dashboard should provide an additional value. For example, a user shouldn’t need to do some more calculations on his own, to get to the information he was looking for, because everything he needs will be clearly displayed on the charts. Always try to put yourself in the user’s position. What data will the user be looking for? What information would help him to better understand the current situation? If you have two relative values, why not add a ratio to show either an evolution or a proportion, to make it even clearer? An important point is also to add the possibility for the user to compare your number with a previous period. You can’t expect all users to remember what were the results for last year’s sales, or last quarter’s retention rate. Adding an evolution ratio and a trend indicator, will add a lot of value to your KPIs and make the user like you.
Good layout choices
Dashboard design principles concern more than just good metrics and well-thought-out charts. Next step is the placement of charts on a dashboard. If your dashboard is visually organized, users will easily find the information they need. Poor layout forces users to think more before they grasp the point, and nobody likes to look for data in a jungle of charts and numbers. The general rule is that the key information should be displayed first – on the top of the screen, upper left-hand corner. There is some scientific wisdom behind this placement – most cultures read their written language from left to right and top to bottom, which means that people intuitively look at the upper-left part of a page first.
Another useful dashboard layout principle is to start with the big picture. The major trend should be visible at a glance. After this revealing first overview, you can proceed with more detailed charts. Remember to group the charts by theme with the comparable metrics placed next to each other. This way, users don’t have to change their mental gears while looking at the dashboard by, for example, jumping from sales data to marketing data, and then again to sales data
1) Avoid scrolling and/or multiple pages for a single dashboard. Anything more than 1 page is considered a report. The graphs and numbers should always be together, allowing the user to do quick analysis.
2) Choose the correct visualization. Take advantage of visualizations and graphs that allow the user to quickly associate patterns.
a. Example 1: Avoid most gauges, what might work for a car’s dashboard won’t necessary work in a business. They take up space and say very little.
b. Example 2: Pie charts with more than 4-6 items can make it hard to compare sizes of the slices.
c. Example 3: Using text instead of a line chart; can give a better understanding of where the data is going
3) Consider flow and transitions. Incorporate links to details and/or filtering, and try to let the eyes rest on the data and graph without unnecessary jumping around.
a. Example: Placing values in the legend but not on the graph, this would cause the user to switch their views constantly when trying to do comparisons.
4) Avoid textual overload, try to keep it visual. Dashboards are meant to be fast and easy to read.
5) Don’t do complex patterns; simple patterns allow the brain to recognize values more quickly. Sort the data to give direction.
6) Group similar characteristics, making them perceived as a group
7) Use the correct grouping. Differing grouping options can convey different messages.
8) Covert complex data into simple logical stories. Arrange items to give a flow of communication with data laid out logically. Are the items in the data distinguishable or can they be removed or combined. Is the macro point being made clearly? Use spacing and subtle borders to create distinctions between groups and add comparative measures, like percentages to communicate differences.
9) Communicate and drive action, don’t try to use every available feature or show off your skills if it is not necessary.
10) Show focus, if the dashboard was designed for a specific reason, then try highlighting that item.
11) Show insight, don’t show metrics by themselves; this will end up leaving the interpretation to the user. Give insight by showing metrics with benchmarks, goals, and prior performance to give context. Give hits & misses, root causes, reference lines, goals, etc.
12) Don’t clutter, always give enough whitespace, and never add background logos, stock photographs, large company logos, or too many links to the dashboard.
13) Make it interactive. Users have their own questions and area of expertise. Allowing the user to filter the view, drill down, and examine underlying data will give them confidence and focus in the area they need most.
14) Stick with the basic everyday graphs, and avoid the cute uncommonly seen graphs. Visualizations like bar graphs, line graphs, heat maps, and scatterplots are popular, because they are easy to read.
15) Keep the write-ups short, sweet and simple avoid dryness. Try to avoid technical jargon, if it makes it less understandable.
SAS BI Dashboard
A SAS BI Dashboard is a container that is nested within a portlet and that contains one or more indicators. An indicator is a composite of one or more related objects. Each indicator has a data source, one or more gauges, hyperlinks to additional information, and range settings for the gauges. Dashboards display critical information in such a way that the information can be interpreted and monitored at a glance. Dashboards can also contain links to other pertinent information, important summary and highlights, and personalized information such as weather, news, and stock news.
Effective dashboard design is both an art and a science. Don’t get caught up in the pizazz, razzle and dazzle of any data visualization just because it is slick, colorful, or you just happen to like the way that it looks. In reality, your creative delivery of information may be confusing, distracting, miscommunicating, or unintentionally undermining dashboard value. Anyone that presents data should strive to make it easy for their target audience to comprehend. A few simple design changes can make a huge difference in context and clarity. Take a look for yourself in this before and after example of not so good and good design.
Not So Good
This “Not So Good” example suffers from a lot of common design mistakes. First of all it has too much information for the audience to quickly understand. The inconsistent Segment colors within charts will likely be confusing. The column chart x-axis quarter labels do not align to the shown monthly values. Single total numbers do not have any context with regards to trends. The pie chart fails to communicate the negative Enterprise results. The Manufacturing Price line chart is way too small and has no legend. The Discount Bands bar chart is also quite small with unnecessary labels. Lastly the table with the scroll bar in the middle of the screen does not add any value. The scroll bar reduces dashboard usability. See if you can spot even more issues with this dashboard design.
A Few Quick Tips
1) Consider Your Audience
Ask how a dashboard will be used and design for next step actions.
What information does the reader need to be successful?
How much detail does the reader need?
What action can be taken and how?
How are exceptions or insights that need action highlighted?
What learned or cultural assumptions may affect design choices?
What do colors mean and can they be visually interpreted?
Which icons are familiar?
Don’t forget to use color blind friendly palettes or icons.
2) Use Best Practice Dashboard Designs
Good design should tell a story with data that does not become overwhelming with way too much information, clutter or noise. Limit content to fit entirely on one screen.
Be cognizant of audience natural textual reading tendency. Starting with the highest level of detail in the upper corner of the screen and show more detail you move down in the direction the audience is used to reading.
Keep your dashboard simple with only a 3 to 5 key values, charts, or tables. Avoid putting too much information on a dashboard.
Remember to provide adequate context and keep related items near each other.
Avoid displaying “singular numbers” without any other context.
Show degrees of change for quick comparisons.
Avoid data visualization variety for the sake of variety.
If detail tables are needed, place them on the bottom of the dashboard
3) Avoid Common Data Visualization Issues
Choose appropriate data visualizations. Don’t use charts that distort reality i.e. 3-D charts. Keep in mind that it is difficult for the human brain to interpret circular shapes. Pie charts, donut charts, gauges and other circular chart types may look pretty but they are not a data visualization best practice.
Be consistent with chart scales on axes, chart dimension ordering and also the colors used for dimension values within charts.
Be sure to encode quantitative data nicely. Don’t exceed three or four numerals when displaying numbers. Display measures to one or two numerals left of the decimal point and scale for thousands or millions i.e. 3.4 million not 3,400,000.
Don’t mix levels of precision and time. Make sure that time frames are well understood. Don’t have one chart that has last month next to filtered charts from a specific month of the year. Don’t mix big and small measures on the same scale, such as on a line or bar chart. For example one measure can be in the millions and the other measure in the thousands. With such a large scale, it would be difficult to see the differences of the measure that is in the thousands.
Don’t clutter your charts with data labels that are not needed. The values in bar charts are usually well understood without displaying the actual number.
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.