Funny how time flies for technological choices. Back in 2015 I wrote about how selecting an analytics dashboard requires deep thought behind business operations to select a useful dashboard. 

That thought today would likely give marketers a headache — there are more well-established data architectures to choose than in 2015. Marketers will select dashboards that incorporate APIs as much as they do from databases, such as SQL. The choices for both are a ‘plenty, compounding an already vast martech ecosystem. Scott Binkler noted the massive size of the market in this post on his marketing technology landscape. Many of these include analytic dashboards. Yet even when marketers solely filter the decision around dashboards only, the options can intimidate.

To reduce the dashboard options into a good choice for measurement needs, let's look at what to consider in developing a dashboard for real time data and analysis.

Quest for Real Time Data and Analysis

As a marketer, let's assume you have sketched out what visual and metrics your dashboard will contain. You have also identified the stakeholders who will rely on the graphs and charts.

Your sketch should help answer — or at least ask — a key question: how does the available data match to the metrics stakeholders need? Does some data require elaborate refinement to align it to the information stakeholders need? The data delivered from SQL and APIs are not always sent in a convenient metric for the audience to understand. That condition implies that you must identify what calculated metrics are affected. Metrics that need to be adjusted quickly will be limited to how much the data must be prepared. 

The second question to ask involves where and how the planned dashboard is regularly accessed. Will the querying access impact the ability to deliver up-to-date material? Access consideration introduces the question of dashboard update frequency versus the frequency of users' needs. The details ultimately dictate the choice of dashboard user controls, how those control manage connection to data sources and having a team assist in upkeeping those underlying connections.

The arrangement of each factor influences the ability for dashboard adjustments on the fly and periodically audit the connections to ensure that the API and database integrations are pulling and rendering the right data to provide stakeholders. Identifying the degree of involvement with data would reveal the amount of maintenance required to avoid technical issues, such as injections of dirty data into a data query stream.

Most tools have a straightforward connection where the data is just imported then appears withing the solution's visualization. These platforms are designed to work straightforward with databases, such as Chartio. It has connectors which allow users to connect to variety of database choices.

Related Article: How Server-Side Tag Management Is Influencing Marketing Analytics

Why Open Source Programming Languages Matter

If you find that you need a myriad of customizations for the data and visualizations stakeholders want, you should look at dashboards supported with open-source programming languages. These programming languages have libraries that introduce graphic user interfaces and custom visualization functions that can be adjusted without an extensive amount of code.

They also connect to a variety of solutions and databases through API token keys, saving developing time in creating a function to access data and send repeated queries for the latest data. There is a learning curve in arranging the code — it's programming, after all — but over the years developers have introduced new ways to make dashboard maintenance easier. This means you can tailor dashboards for individual real-time data needs, simplifying how to arrange the hosting of dashboards and data online.

Most open-source dashboards use R programming or Python as a foundation. Both languages offer statistical and visualization libraries that can quickly model the data. They are perfect for developing complex data models such as regressions, clustering and machine learning models. Thus, a user who is looking a text classification would have several visualizations to support their explanation of the data and conclusions. 

Both languages can also immediately feel intimidating if you're not used to programming syntax. But with some simplicity in coding, you can come up with dashboards that connect to your software solutions through API and update automatically. You can then host those dashboards online to provide the information your team and stakeholders need.

Among R programming professionals, two main dashboard frameworks, Flexdashboard and Shinydashboard, are very popular. They integrate data and visualization using code and syntax in the same way as Markdown, the document format I mention in this post. Like Markdown, users insert a section of code, called a "chunk," which contains the data and graphic code. Flexdashboard can create a series of charts, each visualization derived from its own chunk.

Learning Opportunities


Shinydashboard manages R code and data in a similar way, except that it is designed for shiny apps — apps that run on R programming. Shinydashboard introduces header, sidebar and body functions into the user interface portion of the code.  

shiny dashboard

As a result users can arrange a sequence of charts and relevant comments while using different code, libraries, and data for each chart. The charts contain simple GUI features and can adjust their size for browsers and mobile displays.

Related Article: Google Analytics Introduces Migration Tool for Conversion Transition to GA4

Dashboards Drive Relevant Information for Business Decisions

Many other visualization tools for R and Python provide this flexibility. Developers familiar with Plotly, a visualization solution for Python, are likely aware of  derivatives based on Plotly that enhance Python visualization. One is Dash, a Python framework created by Plotly that allows the user to create interactive data apps and dashboards without requiring a deep knowledge of HTML, CSS or JavaScript.   

Some open-source dashboards introduce integration extensions similar to those on cloud solutions like Google Data Studio. Shinytableau, for example, creates Tableau extensions, shiny apps slightly modified to match R scripts to a Tableau dashboard. So, think of this as a blend of a preferred solution with a familiar GUI (Tableau) with potential statistical prowess added to real-time analysis (R).

To consider how to best refine your dashboard appearance further, read the articles 5 Signs You Need To Update Your Dashboard  and How to Create Dashboard Frameworks That Support Marketing Analysis. Both give ideas for refining dashboard visuals and planning charts.

Dashboards drive the information stakeholders use for their conclusions. By developing the right real-time dashboard capabilities, marketers can manage quick requests from stakeholders much more effectively, noting the potential impact of software and features on analysis meant to enhance decisions influencing the customer experience.