What do you do if your data analysis lacks empathy for the human element?

Com certeza humanizar os dados é o passo mais óbvio a ser executado, afinal, nossas emo??es guiam nossas a??es. Ent?o, é necessário entender a importancia dos dados dentro do negócio em específico, e isso só é possível através de um contato imersivo com cliente, e uma ótima análise exploratória.

Sometimes your negative points in data can be outliers or representation of facts, that do not matter. Imagine you are looking for a car for your business. You found one and you see that some customers rated it with 2 stars. These 2 stars pull the overall review score down. Drilling into the reviews will help you understand the reasons of that issue: imagine all those with 2 stars were not happy with the fact, that the car had petrol engine, which is of no difference to your business. Then, you can remove all those 2-star-ratings and see what is the overall review score then. Same happens in many different cases, this is why we should not solely trust the numbers, but dive deeper to see, what could be a cause of the outliers.

If your data analysis is lacking empathy for the human element you could begin to search for which qualitative data would offset the heavier emphasis on qualitative characteristics of the data. Although we are still dealing with data, which elements of the data are the most human elements of the project? Or which variables can we examine more closely to extract insights that are more personable to our audiences? Perhaps the goal in itself of the data analysis project didn't include a human element necessary for it to advance. In any event, search for where human empathy could be felt or which insights could illicit an emotional response and see if this could be expounded upon in a beneficial way while still adding value to the overall goal.

No meu trabalho sempre procuro contextualizar os dados. Para a maioria dos analistas, uma coluna "salário" dentro de uma base de funcionários, é apenas uma coluna com valores a serem somados e destrinchados na análise. Mas penso que podermos ir um pouco mais longe.

Na maioria das vezes, quando sua análise trata de valores, alguém vai ser "prejudicado". Se eu analiso o valor de vendas total por vendedor, obviamente o vendedor com menor valor de vendas será prejudicado. Dito isso, procuro n?o apenas entregar a análise, mas entregar sugest?es positivas a fim de evitar esses problemas.

On the other hand, you should not be hiding information, that can potentially bring a large scale discussion, spark some debates or challenge varies ideologies and believes. Remember the case of Ignaz Semmelweis, who discovered that washing hands saves lives: at his first attempts to show the medical world the numbers, he was not listened and was laughed at. Sometimes our findings can be against the main flow, in such cases it all depends on the way you present the data. Ignaz was not very good at that and as such, he failed so many times. It is not only important to be able to analyse the data, but also to present it in the right way. The worst is to keep silence only because the data contradicts the accepted and widely practiced knowledge

Sempre toco nesse ponto. Análise que contribue, é a análise que é feita em conjunto a outras pessoas do negócio. Além disso, esses sess?es de "brainstorming" com outras pessoas, consegue nos dar maior dimens?o da importancia daqueles dados e do que precisamos entregar.

Comunica??o é inegociável. Seja com clientes, time, diretoria e etc. Tanto no quesito comunica??o escrita/falada, mas também no quesito de comunica??o visual. Aplicar técnicas de storytelling, pode n?o só salvar sua análise, como elevar o nível dela.