What are some data visualization techniques that can enhance human-centered design?

Understanding your audience is the cornerstone of effective data visualization. Before you create a visualization, think about who will be interacting with your data. Consider their level of expertise, their interests, and the context in which they'll be using the information. Start by understanding your audience - what are their goals, preferences, and pain points? Tailor your visualization to resonate with these users. For example, executives may prefer high-level summaries, while analysts may crave deep, granular insights. Where possible, engage directly with your audience. Listen to their challenges, understand their vocabulary, and identify the aspects of data that matter most to them.

Without a deep understanding of the audience and their genuine needs, no report or dashboard can fulfill its core purpose, which is to genuinely assist users in making informed decisions. Data managers, responsible for crafting these reports in a company, often face the challenge of persuading users to adopt a fresh perspective when interpreting data. In such situations, I've personally employed a side-by-side comparison method to vividly illustrate the most effective approach, even in cases where clients insisted on a different viewpoint. It's worth noting that the level of detail in a report should align with its intended purpose. Strategic reports benefit from simplicity and directness,while operational ones can delve into the specifics.

In my experience knowing your audience is the first thing you need to get right. I have had experiences where the charts and colors and the story were amazing but I lost my audience as they could not interact with the data. Well, that was because I had prepared it in my view and context and completely forgot to include the perspective and experiences of the people who wanted to use the data. So whenever you are visualizing data, STOP and get a read on the audience, it's the only way to ensure that your data visualization adds value to the end user.

In my experience, customizing data visualizations based on the audience is crucial. For example, when creating a sales dashboard for executives, we shall focus on clear charts and metrics that aligned with their strategic goals. Understanding your audience ensures your visualizations resonate, contributing to effective user-centered design.

Data visualization is not just about showcasing data, it’s about making it resonate with your audience. The secret lies in knowing your audience and crafting your design to cater to their needs. For instance, a busy manager might appreciate a dashboard with straightforward charts, distinct labels, and bold colours that emphasise key metrics. Conversely, an infographic aimed at the general public might benefit from imaginative visuals, narrative elements, and interactive features to pique and hold their interest. Therefore, visualisations are there to make the audience engaged in it. After all, the value of data lies in the insights it offers. #DataVisualisation #UserCentricDesign #DataInsights

Choosing the right charts to convey the message that you want your audience to hear is fundamental, a wrong chart can lead to misunderstandings and inability to reach consensus. I usually share the Financial Times Visual Vocabulary PDF with my team prior to starting a visualization project. Another crucial aspect is that in data visualization "less is more" try to keep a dashboard to less than 6 charts/graphs, to prevent a visual overload from your audience and making them spend time deciphering the underlying message.

I understand that the type of data can determine type of chart one is going to use. For example if you are developing a primary health care dashboard for a Country to not only help the key decision makers but also for the health care workers, biostatistician, mid wife's and therefore it will effective if the chart are made simple to use interpret but one that also aligns with what they are used to and you introduce something new proper training is needed which a times may be too much expensive. So from experience especially working in the health sector it always effective to strengthen the existing systems instead of coming with some shinny tools that might never be used, so listen to your users.

Avoid pie charts if you have more than 2 variables. It can be difficult to actual understand each grouping if there are more than 2 - a bar chart is more effective at that point.

Attracting audience towards the data is the key to success. Data should be visually appealing with all needed information. Message has to be crisp and clear with an ease of understanding.

A viz can depict comparisons (e.g. YoY sales comparison), distributions (e.g. distribution of customer footfall in a year), proportions (e.g. revenue contributions by regions), relationships (e.g. relationship between patient demographic and patient mortality rate) and trends (e.g. daily sales trackers). All these categories have a number of different charts and graphs that are ideal to represent the data. There are more case-specific charts like Gantt charts, heat maps, tree maps etc. used in certain scenarios. The end objective should be present information in a way that anyone having contextual knowledge of the data is able to build opinion based on your viz. Keep in mind, using a wrong chart/graph can potentially misrepresent the data.

For all the supporting features of your visualisation (headers, backgrounds, borders, axis, etc.) you should use variations of white, grey and black. This way all the key information is there, but it doesn't distract from the important data. It will also make your end product look refined and professional. Another top tip when it comes to colour is to create a custom colour palette based on your brands colours. There are many colour wheels online when you can enter any hex code and it will give you complimentary colours. By using these colours you will create a dashboard that is aesthetically pleasing but also matches your company's branding.

I have observed one interesting fact about colors in visualization. The business stakeholders who have presentations as part of their day-to-day jobs always prefer similar color schemes in charts. Every organization uses some standard template for presentations similarly a common file can be used across all the business reports to make it easy for both developer's and stakeholder's interpretation. In the case of Tableau, this information can be created once and the .tps file can be shared across for consistency.

Due to my experience in working in a healthcare company, I always tend to use "color-blind" palette offered by Tableau and PowerBI. I believe one can custom build a color blind palette in Excel too. This ensures accessibility across users. Other than this, the best option is make visualizations comply with the company's brand colors.

Humans understand colors in a specific way and we're generally used to certain meanings of colors. The dashboards should usually have a maximum of up to 5 colors to have an appealing design and presentation. Also, this understanding of color may vary in demographics.

Indeed, color stands as one of the paramount and influential elements in the realm of data visualization. Its potency lies not only in its ability to captivate the eye but also in its profound capacity to convey emotions, emphasize critical information, and guide the narrative of the data story. For Example : **Red ** if the cost has increased. **Green ** if the cost has decreased. **Yellow ** if the price is the same.

Having the right context for data in visuals is important as it helps in understanding things better and make informed decisions. Clear visuals simplify tricky and complex info, making it easy to understand. Context always keeps things relevant. Clear visuals also build trust and make it easier to share info. Together, they shape the story in data and help people get important insights to act on it.

One thing that I believe is paramount in really effective visualizations is the effective use of 'white space'. It is the antithesis of clutter allowing the information consumer to best hone in on the insight message being portrayed. Essentially, it's the "less is more" approach to data visualization. What gets measured, gets done. If there is too much going on visually, there is a risk the 'action' message is not received and behaviours don't change.

Data visualizations serve as a means of communication, and like any form of communication, they require a shared understanding between the sender (the creator of the visualization) and the receiver (the user or audience). A simple bar chart devoid of labels, headers, or titles may present information, but it lacks the necessary cues to guide the user's attention or interpret the data correctly.

Según un estudio reciente de Nielsen Norman Group, alrededor del 79% de los usuarios digitalizan contenido web en lugar de leerlo en detalle. Es por esto que, la inclusión de elementos como títulos, leyendas y etiquetas puede ser crucial para atraer y mantener la atención del usuario ( ?~??? ?°)

Dual axes charts usually create more confusion than clarity for the viewer, unless you are trying to illustrate a correlation (positive or negative) between two variables that need to be plotted on different scales. In most other cases, you’re better off generating two separate plots and displaying them side by side instead, so the audience can focus on one thing at a time.

This is only true for exploratory analysis through visualization. Explanatory visuals, should maintain a clear narrative and guide the audience towards the key message while avoiding flexibility and distracting elements. The distinction lies in providing flexibility for exploration versus guiding the audience to a predefined message. Start from a question (exploratory) VS start from an answer (explanatory). So what has been said here is true, just not always ;)

In my experience when the Data visualization product has a large user base, we can't curate it to the needs of everyone. That's when filters, parameters, and actions in the charts play a key role. When the charts have action items rather than filters and parameters in drop-down, users will find it very easy to navigate around and that will truly serve the purpose of building a self-serve tool.

People with years of experience usually like to see the data at multiple levels. It's always better to gather all the data which is available and present it in an interactive way. So they can drill down to the most granular level of data available.

I agree with this point. One major importance of visualization is that it helps the users make sense from a data set within seconds. It is in line with the attention span of the average individual. Therefore, a great viz is one that insight and inference can be deduced from it within seconds of looking at it.

By incorporating interactive elements, one can transform data visualizations into personalized tools that cater to individual interests, promote active engagement, and enhance the overall user experience. This approach not only makes data more accessible but also turns the exploration of information into a collaborative and user-centric journey. Using BI tools such as PowerBI and Tableau enhances interactivity by providing a range of features and functionalities that empower users to explore, analyze, and interpret data in a dynamic and user-centric manner.

In my personal experience, I have learned that applying some agile practices will bring better results when crafting dashboards: 1. Engage your Audience Early: Save time and costs by collecting feedback from your audience as soon as possible. Through an MVP, it is simpler to check if you are on the right path. 2. Sprint Approach: Building on the previous point, your MVP will reveal the next steps to be taken. Try to engage your audience after some progress is made. Your dashboard will grow, fed by its users' reviews. 3. Retrospective: Evaluate what has gone smoothly and identify areas for improvement in your next developments. Lessons learned (as demonstrated in this article) are significant assets for professional growth.

Wir werden demn?chst regelm??ig Dashboard Feedback Runden machen, wo wir darüber diskutieren wollen, wie und wann man Feedback holt und gibt. Intern ist es einfacher, sich Feedback zu holen als extern mit dem Kunden. Die Art der Pr?sentation ist auch wichtig, es hilft, wenn am Anfang Mockups gezeigt werden, die noch keine echten Daten zeigen. Damit verhindert man, dass über einzelne Zahlen diskutiert wird.

Before delving into the data, using post-it notes to clearly outline the story you want to share could greatly help you understand if it's even possible to create the content or if you're missing something. Plus, whatever feedback you receive could be added instantly. It can be risky to dive into the graph creation process without a proper draft beforehand. You might become enchanted by a particular graph and find yourself always trying to incorporate your shiny new graph into the story simply because it took a while to create.

There is nothing more intimidating to the user than an explosion of disparate numbers, shapes, and colors on one page. "Portionizing" amount of information on each page, keeping it clear what question you aim to answer of provide context for, and putting them in an order that creates a cohesive narrative will help user to engage with data effectively and internalize insights.

Prioritising data and charts that provide actionable insights over data that isn't actionable. This way, the value of the dashboard can be measured.

Data visualization plays a crucial role in human-centered design by offering effective ways to present complex information. Techniques such as interactive dashboards, infographics, and storytelling with data enhance user comprehension and engagement. Heatmaps reveal data density patterns, while bubble charts and scatter plots visualize relationships. Treemaps illustrate hierarchical structures, and word clouds provide insights into textual data. Gantt charts aid in project management, and radar charts compare variables across categories. Small multiples enable the comparison of subsets, and 3D visualizations offer immersive experiences when used judiciously.

As BI professional, we need to look at dashboards and reports as products and ensure that they cater to our audience in such a way that each product is intuitive, elements around reading left to right, colour, etc are all considered in design. But we need to consider a template approach to designing our products by developing a template with set features, design, standards and so forth which are common in every product developed. This makes it easier for our analysts to develop as the basic stuff is already done and makes it easier for our users as they don’t need to relearn an entire new product t every time. Further, through testing and user experience being considered, the template will evolve into something even better continuously.

In my recent experience, I've found that data visualizations should be constructed based on users' needs and specific use cases, rather than solely focusing on creating aesthetically pleasing charts. I realized this when encountering a dashboard that had all the necessary data but required a series of steps to extract insights: 1. Selecting a specific number 2. Applying a filter 3. Choosing another number 4. Performing additional calculations Only after these steps did I obtain the desired insights. In contrast, I found it more efficient to download the raw data and manipulate it myself. In this instance, it seems that the visualizations were not chosen optimally.

If I could boil it down to 1 advise... Use interactive dashboards! Example: imagine a dashboard for a retail business that tracks sales data. Instead of using static charts, the dashboard should allow users to interactively filter data by different parameters like months, categories of products, regions, brands ... This empowers users to explore the data in a way that's most relevant to THEM. Many specific needs or questions can be then answered. They could quickly switch from viewing overall sales trends for the past year to analyzing the performance of a specific product category they are interested in, in a particular region over a particular month. This level of interactivity makes the data more engaging and actionable.

Es muy importante conocer a fondo las herramientas de manipulación y representación gráfica, ya que una mala programación o un mal dise?o de enlace puede hacer que fallen los filtros y se muestren datos erróneos o que los gráficos no se modifiquen adecuadamente para representar lo que se quiere mostrar!!!...