RapidCanvas 101: A Quick Start Guide
How to Sign In
Upon enrolling in RapidCanvas, all customers will receive an email containing a link to the sign-in portal. This portal will be used for all future logins. In the email, users will be prompted to create their RapidCanvas account. It is recommended to bookmark the staging login page, as the RapidCanvas homepage does not feature a direct login portal.
Connecting Your Data to RapidCanvas
To connect data to RapidCanvas, follow this simple process to ensure seamless integration from various sources, making the data accessible for AI analysis, AI model and AI solution development:
Select a Data Source Type
Navigate to the Data Connecter section in RapidCanvas and choose the type of data source you want to connect. You will be prompted to add a data connection when creating a new project, and you can add a data connector to your tenant to support any projects within that tenant. Supported data sources include: * Databases (e.g., MySQL, PostgreSQL) * Cloud storage (e.g., AWS S3, Google Cloud) * APIs * Flat files (e.g., CSV, Excel)
Enter Connection Details
For databases or cloud storage, provide the necessary credentials, such as:
Hostname
Database name
Username and password
Security protocols (if applicable)
For files, either upload the file directly or set up a connection to a shared drive.
Validate and Map Data
After establishing the connection, RapidCanvas provides tools to validate and map data fields to ensure they meet the requirements of your analysis or AI models.
Test and Save
Test the connection to confirm that data loads correctly. Once verified, save the connection settings. RapidCanvas will then automatically pull the data for future analysis and processing.
Once connected, your data is ready to be used for building visualizations, training AI models, or conducting further analysis.
What is a Tenant? Creating a Workspace in RapidCanvas
In RapidCanvas, a tenant is a fully isolated workspace designed for a specific organization, team, or project. It ensures data, users, and resources remain separate, providing enhanced security and organization. This isolation enables tenants to maintain strict compliance standards, such as HIPAA, where required.
Key Features of a Tenant:
Custom Settings: Each tenant has unique configurations for data storage, user access, and applications.
Data Security: Data and user activities are isolated to prevent overlap between tenants.
Access Control: Users cannot see or interact with a tenant unless they are explicitly invited.
How to Work in a Tenant:
Access: Log in to your assigned tenant using your RapidCanvas credentials.
Organize: Upload data, create models and solutions, and set team member permissions (if you are a tenant administrator).
Build: Develop AI solutions within your tenant without affecting other projects or models in the platform.
What is a Project?
In RapidCanvas, a project is a structured workspace where users can organize and manage all elements required to build an AI solution, AI Agent or AI analysis. Projects are designed to streamline workflows and centralize resources, making it easier to collaborate, track progress, and deploy solutions effectively.
Key Components of a Project:
Data Integration: Connect and manage data sources relevant to the project.
Model Development: Build and train AI models.
AI Agents: Utilize AI agents to assist in creating models and solutions.
Visualizations: Create and share interactive visualizations for better insights.
Reports: Generate detailed reports to present project outcomes.
Projects keep all relevant resources and workflows in one place, fostering collaboration and ensuring efficiency throughout the AI development lifecycle.
How to Build an AI Solution in RapidCanvas
To build an AI solution, users must first create a project within their designated tenant. Follow these steps to get started:
Create a New Project
Inside the desired tenant, initiate a new project by clicking the purple “+” button in the upper-right corner of the screen.
Name Your Project and Select an Environment
Assign a name to your project and select the environment in which you want to create it. Once this is done, you are ready to begin building your AI solution.
Steps to Build an AI Solution in RapidCanvas
Create a New Project
Start by setting up a project to keep all data, models, and outputs organized in one place.
Connect Your Data
Import or link to the relevant data sources, such as databases, flat files, or cloud storage, ensuring the necessary information is available for analysis.
Prepare and Explore Data
Use AskAI to analyze, clean, and sort your data. This includes checking data quality, removing errors, and uncovering key patterns or trends through AskAI’s generative AI analysis.
Choose Your AI Model
Select from RapidCanvas’s prebuilt AI model options for automation.
Alternatively, create a custom model by defining specific types and leveraging custom-built recipes and parameters.
Train the Model
Run your model on the data to teach it how to make predictions. RapidCanvas, via AskAI, will provide performance metrics and suggest improvements to refine your model.
Test and Refine
Use test data to evaluate the model’s accuracy. Adjust settings, retrain, or fine-tune the model to enhance performance.
Deploy and Automate
Deploy your model to make it accessible to other systems or users. Automate workflows to process new data or trigger actions seamlessly.
Monitor and Optimize
Continuously track the model’s performance, making adjustments as needed to maintain accuracy and relevance over time.
What is a Recipe?
In the context of AI, a recipe is a predefined sequence of steps and data configurations needed to build an AI solution—much like a recipe for cooking a meal. Below, we explain what a recipe does, how to build one, and how to use it in RapidCanvas.
What a Recipe Does
Automates Workflows: Automates repetitive tasks like data cleaning, transformations, model application, and report generation.
Ensures Consistency: Guarantees repeatable and accurate results by standardizing the process.
Simplifies Complex Tasks: Combines multiple actions into one streamlined workflow, reducing manual effort.
How to Build a Recipe in RapidCanvas
Choose Recipe type
Click on the attached data container on your RapidCanvas canvas and select between AI Assisted Recipe, Rapid Model Recipe and API Connector Recipe
AI-Assisted Recipe: Our conversational AI Guide will assist users in building the type of recipe that they need.
RapidModel Recipe: RapidCanvas will auto-generate a recipe, and users can manipulate the recipe by choosing which data to focus on as well as switch between model types
API Connected Recipe: Users can manually connect to an API to generate a recipe inside of RapidCanvas
Edit the recipe
Select each step from available options, such as:
Cleaning data
Selecting a model
Training a model
Evaluating results
Generating visualizations
Arrange these steps in the correct order for your workflow.
Save the Recipe
Save the completed recipe to put the recipe to use within your AI solution or DataApp
How to Use a Recipe
Run the Recipe
Apply the recipe to your data. RapidCanvas will execute the steps automatically in the specified sequence.
Schedule or Automate
Set the recipe to run on a schedule or trigger it based on specific events. This keeps processes up-to-date effortlessly.
Monitor Results
Review outputs such as transformed data, model results, or visualizations. Adjust and rerun the recipe as needed for better results.
If you are unsure about the required ingredients (data) or how to assemble them for your AI solution, RapidCanvas’s AI platform can guide you. It answers questions and provides analysis, helping you design the exact recipe you need for success.
Different Types of Models in RapidCanvas and When to Use Them
Binary Classification Model
A binary classification model predicts whether an outcome falls into one of two categories, such as “yes” or “no.” This is especially useful for making decisions between two options.
Example: Imagine you run an online store and want to predict whether a visitor will make a purchase:
Category 1: The visitor will buy something.
Category 2: The visitor will not buy.
The model analyzes past customer behavior (e.g., time spent on the website, pages visited, or items added to the cart) to predict future customer actions.
How It Works:
The model learns patterns from historical data where the outcome is known.
It uses these patterns to predict whether new customers are likely to buy or not.
Business Use Cases:
Targeted Marketing: Predict which customers are likely to respond to an email campaign.
Fraud Detection: Identify potentially fraudulent transactions.
Automation: Approve or reject credit applications quickly.
Real-World Example: A health app predicts whether a user is likely to stay healthy or get sick based on exercise, diet, and sleep patterns.
Regression Model
A regression model predicts a continuous outcome by analyzing relationships between variables. It answers questions like “How much?” or “What value?”
Example: You run a lemonade stand and want to know how temperature affects sales:
Independent Variable: Temperature.
Dependent Variable: Lemonade sales.
The model analyzes your data and finds a pattern, such as “For every 10°F increase in temperature, lemonade sales increase by 20 cups.”
How It Works:
Identifies relationships between variables.
Uses these patterns to make predictions, such as future sales or expenses.
Business Use Cases:
Forecasting Sales: Estimate revenue based on advertising spend or pricing changes.
Planning: Optimize inventory and staffing based on seasonal trends.
Understanding Relationships: Determine how factors like price cuts affect demand.
Real-World Example: A store uses a regression model to predict revenue changes from increased ad spending.
Multiclass Classification Model
A multiclass classification model helps make predictions when there are more than two possible outcomes.
Example: You own a clothing store and want to recommend an item to a customer. Options include:
Shirts
Pants
Shoes
The model analyzes customer behavior and predicts the most likely category of interest.
How It Works:
Learns patterns from data where categories are predefined.
Uses these patterns to classify new data into one of the multiple categories.
Business Use Cases:
Personalized Recommendations: Suggest specific products based on past purchases.
Organizing Data: Sort customer service tickets into categories like “billing issue” or “technical support.”
Enhanced Customer Experience: Provide tailored suggestions to improve engagement.
Real-World Example: A pizza restaurant predicts the type of pizza a customer will order based on past orders.
Time Series Forecasting Model
A time series forecasting model predicts future trends or events based on historical data over time.
Example: An ice cream shop tracks daily sales and notices higher sales in summer. A time series model predicts future sales trends based on these seasonal patterns.
How It Works:
Analyzes historical patterns (e.g., seasonality, trends).
Uses these insights to predict future data points.
Business Use Cases:
Inventory Management: Forecast demand to avoid overstocking or shortages.
Budget Planning: Predict future revenue and expenses.
Scheduling: Plan staffing levels for busy periods.
Real-World Example: A bakery predicts daily bread sales to optimize production and minimize waste.
Anomaly Detection Model
An anomaly detection model identifies unusual patterns in data that deviate from the norm.
Example: You run an online store and usually receive 100 orders daily. One day, you get 1,000 orders. The model flags this as an anomaly.
How It Works:
Learns what “normal” looks like from historical data.
Flags any data points that deviate significantly from the usual pattern.
Business Use Cases:
Fraud Detection: Spot unusual transactions in banking.
Performance Monitoring: Detect sudden drops in website traffic.
Error Identification: Flag system malfunctions or data inconsistencies.
Real-World Example: A coffee shop notices a sudden drop in daily sales, prompting an investigation into potential issues.
Clustering Model
A clustering model groups similar data points together based on shared characteristics, helping businesses uncover meaningful patterns.
Example: You own a bookstore and want to group customers by their preferences:
Mystery novel readers.
Self-help book enthusiasts.
Parents buying kids’ books.
The model groups customers into clusters, enabling personalized marketing and inventory planning.
How It Works:
Finds similarities in the data.
Groups similar items or people into clusters.
Business Use Cases:
Targeted Marketing: Tailor offers based on customer clusters.
Customer Insights: Understand behaviors and preferences.
Resource Allocation: Focus on high-value customer segments.
Real-World Example: A gym uses clustering to group members based on their preferred activities, like weightlifting, yoga, or swimming, to tailor services.
These models empower businesses to make data-driven decisions, improve operations, and deliver personalized experiences to their customers.
The Difference Between a Recipe and a Model
Recipe
Definition: A recipe is a predefined collection of data points and steps used to create models, analyses, or AI solutions. It acts as a blueprint for building machine learning models by organizing how data is prepared, features are selected, and algorithms are applied.
Example: A recipe might include data points such as a customer’s first name, last name, state, average spending amount, and number of purchases over the past 12 months.
Purpose: Recipes standardize workflows, allowing users to replicate or adjust processes without starting from scratch. They incorporate best practices tailored for specific types of problems.
Model
Definition: A model is the result of applying a recipe to data. It represents learned patterns and relationships from training data and is used for predictions or classifications.
Purpose: Models serve as actionable tools for analyzing new data, offering insights or automating decision-making.
Example: Using the same data, different models—like regression, time series, or clustering—can produce varied types of analyses.
AI Agents: What They Are and How They Work
RapidCanvas AI Agents
Definition: Intelligent, autonomous agents designed to automate machine learning processes and decision-making tasks. They streamline workflows such as data preprocessing, feature selection, and model deployment.
Purpose: Simplify and accelerate AI integration for businesses without requiring deep technical expertise.
Key Features of RapidCanvas AI Agents:
Automation: Handles repetitive tasks like data preprocessing, model training, and evaluation.
Integration: Seamlessly connects with other RapidCanvas components for smooth end-to-end workflows.
Intelligence: Learns from past interactions to improve decision-making.
User-Friendly: Requires minimal setup and is accessible to non-technical users.
Versatility: Adapts to various domains, such as predictive modeling, data analysis, and operational tasks.
Example Use Case:
In e-commerce, an AI Agent could dynamically adjust product prices based on demand or recommend items based on a customer’s previous purchases.
AskAI: The RapidCanvas AI Agent
Definition: AskAI is an interactive chatbot AI Agent embedded within RapidCanvas to assist users in constructing AI solutions, delivering AI analysis on your data and answering real-time queries.
Purpose: Provides insights, explanations, or answers to specific questions without requiring code or manual data analysis.
How It Works: Users ask a question, and AskAI processes it using conversational language to deliver responses based on available data, documents, or knowledge bases.
Example Use Case:
In an e-commerce platform, AskAI could answer questions like, “What was our best-selling product last month?” or “How many customers visited our website today?”
What Is a Job Inside RapidCanvas?
Definition: A job in RapidCanvas is a defined task or process executed within the platform, designed to automate workflows, manage tasks, or process data.
Key Features of Jobs:
Automation: Replaces manual intervention for repetitive tasks (e.g., daily reports or data updates).
Task Execution: Performs specific functions such as data imports, analysis, or interactions with external systems.
Scheduling: Jobs can run at predetermined times or be event-triggered.
Integration: Facilitates seamless interaction with other RapidCanvas components or external APIs.
Monitoring and Reporting: Tracks execution, performance, and outcomes to ensure smooth operation.
Customization: Adapts to user needs by allowing adjustments to parameters and desired outcomes.
Examples:
Data Import Job: Pulls data from external sources and processes it for analysis.
Report Generation Job: Automatically creates and shares reports on metrics like sales performance.
Alert Job: Monitors conditions (e.g., system errors) and sends notifications when thresholds are met.
Summary: Jobs in RapidCanvas streamline operations by automating critical tasks, ensuring consistency, and improving efficiency.
How to Put Your AI Solution to Work
To operationalize an AI solution built within RapidCanvas, follow these steps:
Finalize Your Solution
Review and Test: Validate the model with diverse datasets to ensure accuracy and reliability.
Documentation: Provide clear documentation outlining the solution’s purpose, inputs, outputs, and usage instructions.
Deploy the Model
- Deployment Options: Choose between:
On-Premises: For self-hosted infrastructure.
Cloud Deployment: For cloud-based hosting solutions.
Deployment Tools: Use built-in deployment features, such as a “Deploy” button or guided wizard.
Integration
API Integration: Ensure the solution is accessible via an API, enabling other systems to interact with it.
Data Sources: Connect the solution to real-time data inputs like databases, external APIs, or file systems.
Automate Tasks
Automation Tools: Set schedules for tasks such as batch data processing or real-time predictions.
Jobs: Create jobs within RapidCanvas to automate tasks, triggered by data changes or scheduled intervals.
Monitor Performance
Monitoring Tools: Track metrics such as accuracy, response times, and error rates.
Alerts: Configure notifications for system failures or significant performance drops.
User Training and Support
Training: Educate users on how to interact with the solution, including inputting data and interpreting results.
Help Resources: Provide user manuals or FAQs for quick reference.
Continuous Improvement
Feedback: Gather user feedback to identify areas for enhancement.
Regular Updates: Update the model with new data or algorithms to maintain relevance and effectiveness.
Reporting and Analysis
Performance Reports: Summarize outcomes, highlighting successes and areas for refinement.
Insights: Use findings to inform business decisions, optimize processes, and drive strategy.
By following these steps, you can seamlessly deploy and leverage your AI solution to deliver actionable insights and value.