AI in Action
Curious to see it in action? Watch this video to learn how the AI Assistant works:Step by step to successfully explore your data
1. Know your data
Before deciding which questions to answer, it makes sense to understand what data you have in Nekt, where it comes from, and how they relate to each other. For this, the first step is to check whether the data source you’re looking for is already registered inSources and verify in the Catalog which data is being extracted into the tables.
Use the Preview feature within each table in the Catalog to understand the data format, as this can be important for more effective exploration.
2. Explore your data with SQL
SQL is the most widely used solution for data exploration. In Nekt, it’s the language that allows you to modify and perform operations on the data to answer specific questions. You can use the AI assistant to generate SQL code from natural language commands, or you can create the SQL code yourself if you’re familiar with the language. Let’s take a closer look at both options:Using the AI assistant
Using the AI assistant
The introduction video on this page shows the AI assistant being used and generating results in seconds. You can ask questions in English or Portuguese, whichever you prefer.Our assistant will use the data you have in the catalog and the current SQL query (if there is one) as context. This allows the assistant to generate a new SQL query that answers your questions based on this context. It can even change the table being queried if it determines that the answer to your question is elsewhere (or in a combination of different tables).Generally, you don’t need to worry about the code! Just interact with the assistant until you get the answer you’re looking for. Of course, it’s worth remembering that like any AI assistant, the answers may not always be correct. In this case, you can continue iterating with the AI to correct it, try your hand at SQL to adjust the query, or seek help from someone who knows SQL.
Creating your own SQL code
Creating your own SQL code
If you’re already familiar with SQL or want to try it out, it’s useful to know that Nekt uses Trino as its execution engine. This will help when writing code and knowing which syntax to use and which features (like Common Table Expressions) can be utilized.If needed, you can consult the Trino documentation to help you write specific queries using the correct syntax.
3. Save your queries or create a transformation
After creating the SQL query that produces the expected results, you have several options: 3.1 Save to repeat the query later Use theSave query function to save this query for future use. This will allow you to quickly access this code as many times as you want. Each time you run it, it will be executed on the most up-to-date data. It’s useful when you need an answer frequently.
3.2 Create a transformation and generate a new table
Use the Create transformation function when you want the result of your query to become a new table in your Catalog.
It’s quite common, for example, for a table to come from the data source with more data than necessary for a specific use case or with unformatted data, which makes it harder to use. For this, your SQL code can handle these cases and generate the ideal table - which will only be created and updated in the Catalog (and become available for destinations and visualizations, for example) through a transformation.
3.3 Discard
If you’ve made a query that probably won’t be repeated and whose result doesn’t need to be stored in a table, you can simply close the query and do nothing with it. The Explorer is like a playground - feel free to explore your data and test what results you can achieve.