<aside> 💡 We believe most of the current no-code analytics tools, Spreadsheets, and BI platforms, are not very powerful and business analysts demand much more analytical power without having to learn a complex programming language like Python, SQL, or R.
</aside>
Most data scientists and analysts working in large and small companies these days, work mostly on data preparation, data visualization, and creating simple models, this combination of processes is commonly known as exploratory data analysis. Exploratory data analysis solves most of the business questions asked by people making decisions.
Many of these questions are simple and repetitive. Most of them turn around the main business KPIs to understand the state of business, present, and past, and whether revenue going up or down. These business intelligence questions can be easily answered with pre-defined visualizations generated with dashboards or BI solutions, such as Tableau or Microsoft Power BI, but these simple questions usually lead to more questions that the non-technical cannot answer by themselves (they would require to learn code, SQL, and theory of data visualization ) These tool claim to be self-service but in practice, they are not, they rely on more technical people pre-defining questions for other analysts to consume.
Graphext proactively recommends visualizations as you play with your data, helping you come up with new hypotheses as you validate existing ones. Skip the blank screens and broken charts. Drop in your data and Graphext recommends the best ways to view it.
In order to understand why the KPIs from the dashboards are changing or not and predict what's gonna happen next, one needs to understand the impact of the combination of many variables connected with that KPIs. Traditional BI tools are limited to understanding the relationship between a pair of variables, or three at maximum, but no more than that. The complex patterns, correlations, and models are left to data scientists which work on a completely different territory using code (Python and R mostly) where the business analyst cannot participate.
Data Preparation (also called data cleaning, data transformation, or data wrangling) is normally applied before doing Data Visualization, but very often you will go back to Data Preparation after doing an initial exploration with the visualization. For instance, if you visualize the histogram of the age of your users and notice there are many people who are 99 years old, that probably mean someone decided to codify that way the unknown age of people that didn't fill that value in a form, so you wanna go back to data preparation remove this value and visualize again that chart again. Watch one of these live coding examples from a data scientist to have an idea of the process.