Rath is a augmented Analytic data visualization tool that allows you to automate common analysis tasks and visualize interesting insights simply by importing data sources.
Github: github.com/Kanaries/Ra…
What does Rath do? Why use Rath?
Most of the work of data exploration and Analysis (EDA) is done by humans. Analysts need to understand the characteristics of the data and find underlying patterns and insights. In the past, with tools such as Tableau and PowerBI, these tasks could be quickly done by analysts themselves. However, with the increasing complexity of data, the exploration and analysis of data has become more and more difficult. According to statistics, the number of fields (dimensions, metrics, etc.) involved in a single exploratory analysis data set in a production environment has increased ninefold over the past two years. This means that an analyst who used to have to deal with more than a dozen metrics now has to deal with hundreds. But the analytics tools or BI used by analysts are still using the old manual analytics approach. Rath improves the efficiency of data exploration and analysis by providing automated data analysis capabilities to quickly extract patterns and potential clues from data sets and provide them to users.
In addition, for experienced practitioners who know their industry well but lack the knowledge and techniques of data analysis, Rath can provide intelligent data analysis services that allow business people to focus on the business logic itself, rather than the various analysis techniques and algorithms.
The traditional data exploration and analysis process requires a lot of attempts. Since analysts cannot find fields with potential insights in advance, they need to try and search and locate gradually. This step depends on the analyst having sufficient domain knowledge, data analysis capabilities, and even some visual analysis skills. Rath will help you lower the barrier to use this step, helping you automate statistical analysis and interactive visual design, allowing you to focus on more meaningful activities such as reading, understanding, hypothesizing, and testing. This advantage will be magnified in scenarios with more complex data sets and deeper understanding of the domain.
Data sets such as the Kelper telescope on Kaggle are difficult to analyze without specific domain knowledge. In the face of a large number of fields, analysts often have no way to start. They need to make a lot of attempts before they can find some meaningful rules for analysis. This analysis method is undoubtedly inefficient, and this problem becomes more obvious when the data set is more complex and the number of fields is more. When the data set reaches more than 100 fields, even experienced data scientists have a hard time. Rath provides an automated solution to this problem, allowing machines to automate the analysis and exploration of data sets and provide potentially valuable recommendations to help you find problems faster.
As shown in the figure, it is also difficult to start with convenient visual exploration tools such as Tableau, and it is difficult to find obvious views with statistical conclusions for many dimensions or measurement combinations.
With Rath’s automated analysis capabilities, you can quickly generate visualizations (strong correlations, trends, anomalies, etc.) of statistical underlying conclusions.
As you can see, Rath can sense interesting patterns and patterns in data and automate the design of visualizations to display them. Here is a Demo video of Rath.
(Old version, new version can be directly read below)
We can add down to see the use of Rath steps.
Data source import
First, upload the data set you want to analyze in Rath (currently support CSV, JSON format data), find the file upload button, click. You can choose whether to sample or not (if the data set is too large). Rath supports streaming data for CSV files, so you don’t have to worry if the data set is very large.
After getting the data, you can see the preview of the data in the figure below
After uploading, you can click the Config button to adjust your understanding of the fields (which fields are dimensions/independent variables and which are measures/dependent variables). Rath will help you infer the types of most fields, so you only need to do a general check and adjust the fields that you think are unreasonable.
Or you can use the configuration panel on the right.
In the data source configuration interface, you can also adjust the data cleaning policy and even do data sampling and field processing.
Data analysis and algorithm visualization
When the data source is imported, clickextract insightsButton. The system will automatically jump to the notebook page (old). If you don’t care much about the analysis process and controlling some parameters, you can skip to the Explore or Dashboard interface.
Insight results presentation and visual recommendation
The Explore page displays all the visualizations recommended by the system. You can go through it one by one here. The recommended charts are prioritized, so the charts at the top of the page are recommended more. In addition to helping you find interesting views from large data sets, Rath is here to design more efficient visualizations to help you understand the stories in the data faster.
The number of suicides varies widely among different generations.
The chart above shows that in developing countries with populations between 1,000 and 2,000, the suicide rate for men is significantly higher than for women.
Make associations based on the graph you’re interested in
When you find a chart you’re interested in, click the Association button (the little light bulb) and Rath will help you find the visualizations associated with the visualizations to help you analyze them in more detail.
Associative result presentation
When you are interested in a recommended chart, you can jump to the chart’s page to make subsequent associations based on that chart
As you can see below, the main chart will take you to the page of the chart you were interested in. If you continue to click the Association button, you can continue the association based on the current chart.
New Associative results
If you have a specific purpose and are more interested in a specific dimension or metric, you can use rATH’s search function to search directly for the information you care about:
Insight interpretation: Multiple types of insight findings
Rath now allows you to discover more specific types of insight, such as anomalies, trends, clusters, etc.
In the figure below, Rath will tell you exactly what the recommended views are for, which will explain the significance of the various insight types and provide more detailed information (such as who the specific outlier is) if necessary.
Generate interactive reports/data news directly
If you want to get a more comprehensive look at your data set at once, head over to the Dashboard page and click the Dashboard Generate button to directly generate several visual reports!
This feature could also provide data journalism capabilities. A lot of the time, our daily focus is data-driven, not a set set of dashboards. On a given day, an analyst may be concerned with a number of different dashboards that need to be combined to identify the problem. You can then use RATH’s automated recommendation feature to create a dynamic data news feed that will be automatically delivered to the user each day, providing insight into the real-time data, allowing the user to drill down on the content they care about, or jump to an existing online report.
After obtaining a report, click The Use as Filter button in the upper right corner of a chart to enable linkage
The related resources
Try using the online version of Kanaries Rath Github at github.com/Kanaries/Ra…