Ask Neo

Overview

Neo is an AI agent which has the capability to understand users questions in natural language and provide the required insights based on the user’s data stored in their data warehouse and made available to Neo for analysis. This allows all users, irrespective of their knowledge of SQL or data warehouses or any other programming language like Python, to perform data analysis on such large data sets. In order to do this Neo needs access to the following:-

  1. Your data, which should be in the data warehouse and ideally stored in a manner suitable for analysis (this can be achieved easily using Transformation features of DataChannel. Click here to know more).

  2. Knowledge of your company and business terms etc along with the description of various columns that your data may contain. This is a part which will be refined more and more as you interact with Neo.

How it works

There are three major parts to Neo:

  • Collections- These are the knowledge base for Neo and are used to answer all the questions asked by the users. Collections are a set of related tables that are made available to Neo (along with their descriptions and any other information that Neo may require) to be able to carry out analysis. The more we refine this knowledge base, the better would be the quality of insights that Neo will be able to provide.

  • Chat- As of now, the primary way for users to interact with Neo is through the Chat interface. This can be used to ask Neo for insights on the data made available to it through collections. This is also the place where Neo provides the insights in the form of text answers and supporting charts and visualizations where required. The users can also observe the thought process used by Neo to arrive at the insights and review / copy the exact SQL queries written by Neo to carry out the analysis.

  • Sources- Sources are the databases, data warehouses, or flat files that provide the data used by Neo for analysis. You can connect supported data sources such as Amazon Redshift, Snowflake, PostgreSQL, Azure Synapse, Microsoft SQL Server, MySQL, ClickHouse, Google BigQuery, and Flat Files. Once a source is connected, its data can be used to create Collections, enabling Neo to query the data and generate insights. A source can contain multiple databases, schemas, or tables depending on the selected data platform.

TIPS

Here are some tips for getting more reliable results in response to your queries:

  1. Give clear table and column descriptions. Also, ensure that there are no missing descriptions. This is essential for AI to generate meaningful insights and accurate SQL Queries.

  2. Metrics and dimensions (that are existing as column fields or are to be calculated) should be clearly defined. Give a calculation formula (based on existing fields) wherever possible.

  3. Keep in mind that these models are still in the nascent stage and hence complex relationship hierarchies can not be figured out by AI without human intervention. Keeping in mind these limitations, it is advisable to break down complex problems into smaller sub-problems.

  4. Do not forget to give a like for an accurate answer and a dislike for an inaccurate one. Ask Neo depends on your feedback to get trained for future interactions. This process is an iteration of learning, unlearning and relearning.

  5. Always make sure to make the query as clearly worded as possible. Ambiguous queries will lead to inaccurate results.

  6. It is important to ask follow up questions to refine your results and drill down to get deeper insights.

  7. It is important to verify/ validate the results obtained using AI. Know that Ask Neo can only assist you by making your work easier. It can not replace you. Without detailed explanations, Ask Neo is likely to struggle with analysis of nuanced or uncommon use cases/ edge cases.

  8. A detailed description of your company and the collection will help Neo to understand the context of the queries better. This will help in generating better results/ recommendations for further conversations.

Overview of starting with Neo

Familiarising with Collections

Click here to understand collections better.

Creating a new Collection

The step-by-step process outlined here will guide you through the process of creating a new collection.

Editing an existing Collection

You may edit and make changes to an existing collection, by following the steps listed here.

Finding an existing Collection

Navigate to any desired collection using these simple steps.

Deleting an existing Collection

You may delete any collection using these simple steps.

Managing a Collection

Once you have created a collection, you’ll need to manage users having access to it. Just follow the steps outlined here.

Familiarising with Chat

Click here to acquaint yourself with the user interface of the chat.

Creating a new conversation

The step-by-step process outlined here will guide you through the process of creating a new conversation with Neo.

Finding an existing Conversation

You may navigate to an existing conversation, by following the steps listed here.

Exploring data Sources

Click here to understand sources better.

Connecting to Amazon Redshift

Follow the step-by-step guide outlined here to connect your Amazon Redshift data warehouse to Neo.

Connecting to Snowflake

Follow the step-by-step guide outlined here to connect your Snowflake data warehouse to Neo.

Connecting to PostgreSQL

Follow the step-by-step guide outlined here to connect your PostgreSQL database to Neo.

Connecting to Azure Synapse Analytics

Follow the step-by-step guide outlined here to connect your Azure Synapse Analytics workspace to Neo.

Connecting to Microsoft SQL Server

Follow the step-by-step guide outlined here to connect your Microsoft SQL Server database to Neo.

Connecting to MySQL

Follow the step-by-step guide outlined here to connect your MySQL database to Neo.

Connecting to ClickHouse

Follow the step-by-step guide outlined here to connect your ClickHouse database to Neo.

Connecting to Google BigQuery

Follow the step-by-step guide outlined here to connect your Google BigQuery project to Neo.

Connecting to Flat Files

Follow the step-by-step guide outlined here to upload and connect Flat Files to Neo.

Still have Questions?

We’ll be happy to help you with any questions you might have! Send us an email at info@datachannel.co.

Subscribe to our Newsletter for latest updates at DataChannel.