Early this year, as part of a project which includes an exciting range of new Javascript based apps called Invotra Labs, Invotra developed an admin dashboard app. This app includes intranet metrics which can give Intranet Managers a quick snapshot of specific intranet activity over a specified time.
After deciding what intranet data sets to display and which graph and chart solution to use, we proceeded to build out the dashboard screens.
Some initial research was done to understand best practice when using charts and graphs and building a UI framework of chart and graph types and components. This work was helpful in making better-informed decisions when building dashboard screens.
With this new found knowledge, I investigated further to fully understand beyond a UX or UI designer’s perspective the importance of data sets and how they are best visualised.
So, why are visualised data sets important?
In today’s digital world with access to ever-growing data sets that are publicly available or generated by large organisations, there is an ever growing problem of what to do with that data.
Big data can remain just that, worthless unless it can somehow tell us a story and therefore offer us insights we can use within our business.
Visualised data helps with business intelligence including:
- Business analysis of the large data sets that your intranet can provide
- Making key strategic decisions based on analysis of this data
- Making information easily understood
- Pinpointing emerging trends
- Identifying relationships and patterns
- Communicating data stories to others
People in your organisation that benefit from dashboard metrics include:
- Webmasters
- Internal communications
- IT Managers
- Analytics leads
- Section managers
- Publishers/Editors
- Organisational users
Before you get started… set requirements or state a purpose
Before jumping right in and building a collection of dashboards it is useful to define the requirements or state the purpose of each screen or set of screens.
This can be broken down into three key considerations:
- For whom are you doing the visualisation (Your audience)
- Scope (What type of input data do you want to include?)
- Outcome (What actions are your audience going to take?)
An example of a purpose statement:
To provide tools to intranet managers in order to help them prioritise their goals for promoting intranet engagement by new and existing users by visualising the top 5 KPIs in charts or graphs that show trends over time for the past year of all social apps and user profile activity.
Steps of data visualisation
Whilst this article is primarily concerned with the visual and interactive aspects of data and dashboards, it helps to have an understanding of the other steps involved in delivering a set of dashboard screens.
There are two main approaches to Big data:
- Data mining
- Ingest a preset data source
- Clean or munge it
- Mine it
- Visualise a conclusion
- Real-time processing (this is the method Invotra are using)
- Capture data streams
- Store it in an appropriate format (potentially munging it with other data as you store it)
- Process data in the background or on the fly as requests come in
- Create either preset visualisations or use a dashboard tool like PowerBI or Google Data studio to connect to an API.
The three main categories of data visualisation
When looking at the composition of a dashboard screen there are three types of data visualisation that can be used, either in isolation or together in groups.
These are:
Strategic
- Display large data sets over a specified time period (Analysing trends over time)
- Can feature relationships
- Can show comparisons
- These are useful in setting business goals or new strategies
- These can reveal why an action took place
- These require in-depth analysis
Useful for
- Allows users (high-level management for example) to make strategic decisions. For example, are more people using the Intranet search function over a specific time?
Visual representation
Time-based graphs, charts or tables that show:
- Time series
- Ranking
- Part to whole
- Deviation
- Distribution
- Correlation
Operational
- Displays real or near real-time data sets
- Most KPI (key performance indicators) dashboards fall under this category
- KPI values below or above certain levels
Useful for
Allow users to understand current business activity and whether immediate or near future action or responses is required. Such as discussion/comment threads that are becoming hot in the organisation.
Visual representation
Real-time or near real-time. KPI modules, stacked typographic devices
Single stat headline or impact number.
Tactical
- These offer contextual analysis
- This is a mixture of strategic and operational data sets
- Offers just enough information to tackle an issue or issues
Useful for
Guiding users to take a decision which may require some further analysis
For example, seeing which team has posted the least messages of thanks or gratitude in the organisation.
Visual representation:
Combination of Strategic and Operational types
Refining
Visualised data sets in the form of charts and graphics can be further enhanced by adding interactivity and animation.
This not only brings the data displayed to life but also allows you to organise multiple views of a specific graph or chart without having to add additional versions on the same screen.
This is simply done by using:
- A toggle view
- A hover option over the data points
- Filtering or sorting option
- Click, tap or keyboard interactions
Visual data should also be accessible on all devices and for user abilities.
Remember… It’s all about the data stories
The quality of the data sets and how they are visually represented should be your main focus when building dashboard metrics, and while aesthetic or interactive attributes are important, these should be secondary objectives.
Dashboards can consist of one or more of the three categories defined above.
When this occurs it not uncommon to have a core feature or features between these categories.
Core features are defined as those features in the visualised data on a dashboard screen that offer maximum impact, or that highlight key insights to drive business decisions or actions. (Pareto principle)
Beyond features and functionality
At Invotra, as we continue along the path of building dashboards and metrics for large enterprise intranets, we are looking beyond features and functionality. We want to build experiences where business leaders and managers can harness the powerful insights presented in their visualised intranet data to make impactful, lasting changes and effective, strategic decisions within their organisations.
Further reading:
Gov.uk piece about the performance of government services
A useful collection of blogs about “data in government“
An insightful blog into the data visualisation process
A javascript library for manipulating documents based on data
If you are interested in a visualisation which was used to change the institutions of London (and save peoples lives) watch this video.
