Data analysis plays an important role in today’s data-driven world, where businesses and organizations collect vast amounts of data in order to make informed decisions. The role of a data analyst is to extract insights from this data, which can then be used to improve business performance and drive growth. This involves analyzing large amounts of data in order to identify patterns and trends to draw meaningful conclusions that can be used to inform business decisions. For example, a data analyst might analyze customer data to identify the most popular products, or to understand customer behavior in order to optimize marketing strategies.
How to do Self-Service Data Analysis in DvSum?
Watch this quick video tutorial on how to do Data Analysis in DvSum app.
However, you can also follow the steps mentioned below which will help you in analyzing the data.
DvSum DI Catalog has a unique and powerful feature through which any user can perform Self Analysis on their data. Currently, this feature is enabled for Snowflake, Databricks, PostgreSQL, and Azure Synapse catalogs. Go to the table detail of any of these mentioned sources and select the Analysis tab.
From the Analysis tab click on the "New Analysis" button and a new analysis will be created.
In the newly created analysis, the user will see a default question section which is also known as NCQ (no code question) which shows sample data of the table by default. On the bottom there will be interactive chat bar where user can ask different type of questions from the data present in added Tables for the purpose of analysis.
In the query block, click on the Profiling icon. A profiling section will open which will show all the profiling of each column. Click on each column and the visual distribution of that column will open up.
From the profiling, the user can select any column having a datatype of integer, string, or date and select some filter accordingly and apply it to the data. Once the filter is applied, sample data will be filtered.
After applying the filter, run the query and it will fetch the results from the real data in the database according to the applied filter.
Multiple filters can be selected and applied from the profiling section and results will show up accordingly. So that's how users can extract insights from the data.
Conversational AI for Data-driven Insights:
One of the key features of this Self Service Analysis is conversational data-driven insights. Yes, the DvSum Data Catalog app has the capability to allow a user to ask a question from its data catalog for precise data insights and analysis.
On the bottom of every analysis there will be interactive chat bar where user can ask different questions and get useful insights from the data.
But before asking a question, one has to set the context from where the question will be answered. For that click on Pinned tables in the Explorer section which will show all tables which are pinned. Remember that we can set the context for the question in two ways:
- Pinning the Tables
- Adding the Context from the Question block
If the user is interested in asking the question from the whole table then user can pin that table and it will be added as a context for the questions but if user either wants to add whole table or add some columns as context then he can add them from the Question block. Both the methods are explained below.
In the example given below 2 tables are pinned, which means questions will be answered from the 2 pinned tables.
How to Pin a Table?
Click on "Table Browser" from the Explorer section. Hover over the tables you want to pin and click the Pin icon. Once the table is pinned, it will also appear in the "Pinned Tables" section. If that table is added for the analysis then that table will be automatically added in the pinned section.
As shown above, 2 tables are pinned. Now whatever question is asked, will be answered by keeping a context from these 2 tables. Below is an example of how a question can be asked and how its results and insights will show.
Users can also examine the query it generated based on the question. Click on the "Show Code" icon, and it will show the code view of the query that was generated,
Saving Context Manually
Context for any table or any specific filter applied on the table can be added from the Default Question block:
The filter that will be applied and and its context is saved then this added context can be seen below the table. User can add as many as contexts either from same table or different tables. Next Time if any question will be asked then this context will be remembered. If the context is no longer needed then it can be removed from Clear Context.
Now if same question that was asked above is again asked but with the context added then answer will be generated according to the added context and it can also be verified from query code:
Here one thing should be remembered that the context which is saved from the question block has the precedence over the context that was added when some table is pinned.
If the analysis is completed then the user can see summary of the overall chat. Here one thing should be noted that Summary will only be generated when statistical questions which involve keywords like average, sum, min, max etc. are used:
DvSum's Intelligent Conversational AI will not only generate Insights about the Data but it will also suggest some recommendations based on the generated insights thus helping in making better decision about the data:
So that's how Conversational Data-driven Analysis can be used by any user to get more meaningful insights into the data.
Sharing Analysis with other Users:
When the Analysis is saved, the "Actions" button will appear on the top right of the screen. Click it and select the "Share" Analysis button.
A Share modal will appear which will show two options for sharing:
1. Everyone: All the users will be able to see the Analysis.
2. Specific Group or People: A specific user or a group of users can be selected with this option
Once it is shared, the selected users will get an email notification and they will be able to see the Analysis.
Note: The user with whom some Analysis is shared, will just be able to view the Analysis and won't be able to make some changes to it. If the user wants to make some changes in the shared Analysis, then first he needs to duplicate that Analysis and then make changes in the cloned Analysis.
Duplicating an Analysis:
When the "Actions" button is clicked, there is an option of Duplicating an Analysis. Once the user clicks on that, a new duplicate of the existing analysis will be created. Users can make changes in the cloned Analysis and save it.