What are the most common data visualization mistakes made by Business Intelligence (BI) professionals?

Sometimes the visualisation which is easy to understand for the BI expert may not be easy for the audience. So though the chart we choose as a BI expert could be right for the data, it may still not be fit for the audience.

Common data visualization mistakes made by BI professionals include: Overcrowding: Including too much information in a single visualization, leading to clutter and confusion. Misleading Representation: Choosing inappropriate chart types or scales that distort the data and mislead viewers. Ignoring Audience: Failing to consider the audience's knowledge and needs. Lack of Context: Presenting data without providing sufficient context , making it difficult for viewers to interpret. Poor Labeling and Formatting: Neglecting clear labeling of axes, making it challenging for viewers. Failure to Tell a Story: Presenting data without a clear narrative , missing the opportunity to convey insights effectively.

One common data visualisation mistake made by Business Intelligence (BI) professionals is failing to verify their assumptions. This can lead to misleading or inaccurate interpretations of data, as assumptions about the data's structure, quality, or relevance may not always be correct. Without thorough verification, BI engineers risk producing visualisations that misrepresent key insights or fail to address the intended business objectives.

Para iniciar a constru??o de um dashboard, primeiramente você precisa saber o que ele precisa te responder. Após, escolher os gráficos corretos para demostrar com clareza as informa??es. A escolha de gráficos errados, podem distorcer a informa??o. Procure manter um dashboard limpo e claro para n?o confundir o público com muitas informa??es.

Uma visualiza??o errada compromete totalmente a análise a ser alcan?ada, antes de criar qualquer gráfico estipule quais métricas e KPIs a serem analisados, após analise as dimens?es que rodeiam essas analises, a quantidade e os tipos de dimens?es envolvidos ir?o nortear quais visualiza??es poder?o ser usadas para ter sucesso na análise. Para n?o ocorrer de realizar muitas análises em uma mesma pagina de Dashboard é essencial dividir em 3 grupos: Análise Estratégica, Análise Tática e Análise Operacional.

The "rule of 3" applies to data visualization; Charts that include more than three concepts are difficult to quickly understand and remember.

Use graphics that communicate well with your audience. In many cases, avoid extremely technical information. Use big numbers to show good results and alarming information. Use few graphics to show detractors, promoters, and what must be done to achieve the proposed goal

Le r?le de la data visualisation est de présenter de fa?on claire les données aux décisionnaires. Il est par conséquent important, après avoir choisi les bonnes visualisations, de pouvoir les agencer de manière à ce qu'elles renvoient l'information souhaitée et aident à la prise de décision.

Overloading data visualizations is a common mistake that BI (Business Intelligence) developers must avoid. Too much information in a single visualization can be confusing and cannot deliver the message. Instead of cramming everything into one complex chart, identify key points to communicate and create simple, related visualizations. This will deliver the message and the visualizations will be cisp and clear.

Common data visualization mistakes by BI professionals regarding data overload include: 1. Cramming too much information into a single visualization, leading to cluttered and overwhelming charts. 2. Not prioritizing the most relevant data, causing important insights to get lost in the noise. 3. Using overly complex visuals that confuse rather than clarify insights. 4. Failing to provide adequate interactivity or drill-down options for users to explore detailed data points. 5. Ignoring the principle of simplicity, which can hinder audience understanding and engagement.

No matter how insightful the analysis, it must be adapted to the stakeholders who will consume it. Always start by considering the audience's perspective and what data stories will resonate before building out any visualizations. Ignoring context and audience leads to poor communication and engagement. Also, always highlight key takeaways and strategic insights that would resonate with them.

Understand the business outcome of the visualization FIRST before painting the visualization. Technologists are too quick to begin developing before understanding the business goals.

Ignorar o contexto pode levar a uma falta de relevancia na apresenta??o. N?o considerar o conhecimento prévio ou nível de compreens?o do público pode levar a uma apresenta??o que é muito complexa ou muito básica, é isso pode resultar em um desinteresse por parte da audiência.

Além de conhecer o nível de conhecimento e o contexto do público é interessante usar exemplos do dia a dia dos expectadores, além contar história com os dados você cria uma conex?o real com o público.

Se seu BI tiver "big numbers", ou seja, cart?es com indicadores oriundos de um cálculo, para ficar ainda mais claro, uma dica é colocar a memória desse cálculo, sinalizando como você chegou naquele resultado. Ex.: Custo per capita. Colocar algo que explique o cálculo, como "total de custos / vidas distintas".

No quesito de Uso indevido ou manipula??o dos dados é essencial prezar por fontes de dados confiáveis e estáveis, na minha opini?o fontes manuais devem ser evitadas ao máximo, como Planilhas Excel por exemplo. Nesse sentido o melhor a se utilizar s?o Banco de Dados.

Accurate data representation, including the use of appropriate graphs and measurements, is essential. Misleading visualizations can sway stakeholders, prompting decisions built on false assumptions and potential financial repercussions. An example is a bar chart where the y-axis begins from a non-zero value, magnifying differences between variables inaccurately. Such manipulation can severely damage the analyst's credibility, which is a loss that's hard to mend.

Due to pressure to present “good” data, the team may misuse and manipulate data to convey a convincing story to stakeholders/ shareholders involved. Cross checking across reliable sources is very important.

Ethical Guidelines: Establish and adhere to clear ethical guidelines for data analysis within your organization. Clearly communicate the importance of unbiased and accurate reporting. Transparent Reporting: Practice transparent reporting by clearly detailing the methodologies, assumptions, and limitations of your analysis. This helps stakeholders understand the context and potential biases. Data Governance: Implement robust data governance practices to ensure the quality, accuracy, and integrity of the data used in analyses. Regularly audit and monitor data sources.

An example I have been facing is the accuracy of data transformation, you have to be very cautious when you apply cleaning steps, specially with duplications.

BI professionals sometimes focus too much on their own interests, neglecting whether the visualization effectively communicates to the audience. This can lead to visuals that are too technical,simple or not aligned with the audience's knowledge and preferences. To avoid this, understanding the audience's level of expertise, interests, and context is crucial. Tailoring the visualization accordingly ensures it's engaging, informative, and relevant.

Sobre o quesito "Esquecendo a história e a mensagem", é muito arriscado durante o desenvolvimento de um Dashboard que você se perca no objetivo principal e acabe misturando contextos. Para mitigar esse problema recomendo sempre navegar por todo o Dashboard cruzando o máximo de filtros possíveis, assim caso haja uma visualiza??o desconexa você irá perceber.

Algo útil é iniciar com uma história breve, que gere interesse e curiosidade ao público. é importante que a história tenha come?o, meio e fim, com alguns detalhes que permitam a visualiza??o mental da narrativa e entrega da mensagem.

Often Data Analysts get lost in the technicalities/detail of the data loosing sight of the fact that it's not the data the audience wants to see, but the story behind the data. In constructing the visuals it's imperative to understand who will be looking at them and what message do they need to be told.

It’s easy to get carried away and go down rabbit holes. Stick to your requirements and if there are valuable insights you find along the way, share it with deviating from the core requirements.

Something very simple but overlooked many times is inadequate data labeling. This can create confusion and tweak the story you are trying to tell. Ensure that axes, legends and data points are clearly labeled to help end user’s understand the right context. Another important one that gets overlooked is ignoring the audience. This can stem from poorly conducted discovery sessions. The most important question to ask is - Personas and how are they going to interact with the data/visualizations. You need to tailor your visualizations to the target audience. Use correct terminology and context that will resonate with your audience. This will definitely help with adoption.

Visualisation should contain data for decision making not for analysis. Keep it simple and precise. Ignoring context, data quality and mobility responsiveness.

Not telling the story and build up the visual incrementally, thus allowing inclusion and exclusion of dimensions that make the rest difficult to interpret.

Finally, a decent question. There are many visualization mistakes that BI professionals make. At the top is chart type. Most chart types are fairly useless in visual analysis. The only ones that really work, besides just a data table, are simple representations such as line and bar charts. Pie charts are the most notorious for providing deceptive visual trends. Horizontal bar charts are more effective than vertical bars, especially for small differences. Gauge charts are an example of something gimmicky - use bullet charts instead. Over usage of bright colors is another one - doesn’t really help - but we do it because bright colors are attractive. Best to just read Tufte etc, internalize their principles, and start applying them.

Keep your visualizations in a way that they hold attention and deliver the necessary message. A crisp and simple visualization may also lead to a significant impact.