Jump to content
  • SDG Group


    This article provides some best practices for creating Spotfire applications

     

    About us

    Consultancy company Halpenfield has become part of SDG Group.

    Our core skills and knowledge in Spotfire means we're uniquely placed to solve complex business problems in industries increasingly defined and disrupted by data.

    With our experienced Spotfire consultants, data scientists and technologists our team has the capability and creativity to analyse complex business challenges, define original solutions and build bespoke tools that improve business performance.

    We're passionate about cutting-edge technologies like Spotfire, but it's only ever an enabler. Everyone at Halpenfield is a business problem-solver.


    Visualisation / Dashboard Best Practice

    Why / How?

    Building visualisations & dashboards as a way of communication have been all the rage for a long time now and for a good reason. They can help you and your coworkers build a collaborative environment to share knowledge and gain insight from one of your companies most important assets - data.

    We all know that they can help your organisation get on the same page, make faster decisions, drive operational / financial efficiencies and find revenue. What's more everyone is looking at a single consistent view of the business to help create organisational alignment.

    So yes they are effective but how can you make sure your Spotfire dashboards are working as they were intended? Are you asking the rights questions to your audience? Where do you start? What's the best visual? Is it going to deliver the right information to help the company?

    At HAL we have a proven methodology that can deliver focused, stunning Spotfire Dashboards that flow and will provide your organisation with the information it requires.

    What next

    To follow is a highlight of the best practices to take:

    1. Always work with manageable chunks  be agile in your approach. Make sure you are able to deliver a ?live? dashboard in weeks not months.
    2. Identify the right metrics to present  Make sure they are necessary / contribute to the objectives & they can be accurately delivered from your internal / external data.
    3. Make it visual  Humans are much better as processing pictures than numbers. Think about the right visuals ? Each visual has its purpose i.e. Line chart for a trend, scatterplot for outliers. Go easy on the pies!! Stick to this and you won?t lose your audience
    4. Draw your audience by making it interactive  You dashboard needs to tell many stories for different users. This is the key difference between a reporting dashboard and an analytical dashboard. Provide a range of filters and drill downs. This will help look at the big picture where most people need to start but then drill down to the view that gives them the information they need to get the job jone
    5. Provide context  Visuals can be meaningless unless they can be measured against something e.g. this year vs last year, actual vs target.
    6. Keep it updated / evolved - If you don't its not worth having it at all! Metrics need to be updated to reflect current business challenges. You need to keep up with the users when their analytical needs evolve.
    7. Keep it simple, focused & easy to access  Masters of visualization such as Few & Tufte always stress the importance of ?white space?. Give your visuals breathing space to avoid confusion.

    Applying this to Spotfire

    Since Spotfire version 7 we have been given added capability on how the application can be customised by editing the theme. Utilising this along with various tricks and techniques of cleaning your application and focussing on the right metrics you can provide your audience with luring dasboards.

    Lets take the simple example of Expense Analysis. We need to measure expenses by department by expense type, show a trend and provide some underlying figures

    Before:

    expensesbefore_0.thumb.png.2a9427c923660a924d2d1c45c2ea44d7.png

    After:

    expensesafter_0.thumb.png.b2f9cb59ee7559abdfa0b7351c300f34.png

    Tips to achieve this style:

    • Soften the fonts and background colours using "Themes". Dark grey font colour is easier on the eye
    • Remove borders from individual visualisations in the theme
    • Build your own navigation and filter panels using text areas
    • Minimise the legend down to what's absolutely necessary ? of you are using the same colours on other visuals then only display once
    • Remove Category / Value Axis Selectors
    • Remove unnecessary scale labels. If you are showing the label then hide the scale. If you are not showing the label then sown the scale.
    • Use the visual as the filter  make use of details visualisations. It might remove the need for filter panels entirely
    • Group visuals together  this will help guide the user how you want them to interpret you want to convey
    • Only show what's required e.g. Filter down to specific department. Show Top N expense types.
    • When using crosstabs highlight areas of concern / importance
    • Think about the "white space"  provide breathing space for what's being presented. i.e. splitting line charts into smaller spark lines can often be easier to read
    • Be consistent on font choice  Segoe UI is a current fav! Think like a newspaper start headings big then reduce down into the detail
    • ....and finally. Learn how and when to use the box plot!!

    boxplot2.png.7dd1b7c9a91ae18dca838009e8e8ca88.png


    Extending Spotfire

    There will always be a need to extend software capability beyond "out of the box" and spotfire grants us several ways in which to do this:

    HTML, CSS and Javascript customization

    - Changing the default look and feel of text areas, or making text areas respond to user interaction; perhaps to dynamically swap out spotfire filters.

                               dynamic_filter_0.png.2b0e6500008de28e003c8d7fc3f80e79.png

    Custom vizualisations using JSViz

    - For example, to plug in a vizualisation otherwise unsupported, such as a highcharts radar.

                             jsvis_example.png.9f17e7bc325200c5ef95e88286d96013.png

    Iron python scripts using the spotfire SDK

    - For easy and clean manipulation of charts, or to explicitly control when and how data is reloaded.

    ip.png.f87e0d3b3c1228577a878315d8eca0d6.png

    .net plugins using the spotfire SDK, deployed as packages.

    - For example, to create a brand new export utility enabling MS Office capability against spotfire vizualitations with fully editable data (powerpoint, excel etc).

    ppt.png.e49c59541a0d92326376fe48b9921741.png

     these are all real world examples where we have used our technical expertise to deliver above and beyond the default capabilities offered in an already powerful BI tool.


    How To

    We have produced tutorials and examples of how we've overcome various spotfire challenges - find them on our "How To" page.


    Contact Us

    We hope you find this page useful. Do you have a challenge we could help with? Get in touch!

    SDG Group in the Spotfire Partner Locator


    Attachments

     

     


    User Feedback

    Recommended Comments

    There are no comments to display.


×
×
  • Create New...