# Pecan Help Center

> Got a tough ML nut to crack? We're here to help!

## Getting Started

- [Why the Right Predictive Question is Key to AI Success](https://help.pecan.ai/en/articles/9217358-why-the-right-predictive-question-is-key-to-ai-success.md): Pecan's Predictive GenAI simplifies the process, turning complex AI modeling into intuitive steps. Dive in and propel forward!
- [Building Your Own ML Model with Pecan: A Walkthrough for Everyone](https://help.pecan.ai/en/articles/8300622-building-your-own-ml-model-with-pecan-a-walkthrough-for-everyone.md): Unlock the power of machine learning in 4 easy steps with Pecan—no PhD required! Perfect for BI analysts and data enthusiasts alike.
- [Getting started: What table you need for predictive success](https://help.pecan.ai/en/articles/8232509-getting-started-what-table-you-need-for-predictive-success.md): Equip Pecan with your transactional table and unlock powerful insights. Your data's future starts here!
- [How can I contact Pecan support?](https://help.pecan.ai/en/articles/6722360-how-can-i-contact-pecan-support.md)
- [Pecan's demo data](https://help.pecan.ai/en/articles/7119621-pecan-s-demo-data.md)
- [Best Practices for Your First Notebook](https://help.pecan.ai/en/articles/10808928-best-practices-for-your-first-notebook.md): Kickstart your first model with a simple target, an event-based trigger, minimal attributes—grow gradually and refine iteratively.
- [Pecan's Data Science: A Peek Behind The Scenes](https://help.pecan.ai/en/articles/8269440-pecan-s-data-science-a-peek-behind-the-scenes.md): Learn about Pecan's state-of-the-art pipeline for tailored ML models and automated feature engineering - while keeping your data safe.
- [Binary vs. multiclass vs. regression models](https://help.pecan.ai/en/articles/6549974-binary-vs-multiclass-vs-regression-models.md)
- [Three types of machine learning](https://help.pecan.ai/en/articles/6549972-three-types-of-machine-learning.md)
- [Pecan & Information Security](https://help.pecan.ai/en/articles/8182379-pecan-information-security.md): Secure your insights with Pecan! Learn how we protect your data and maintain robust information security.
- [Pecan's Glossary](https://help.pecan.ai/en/articles/7850058-pecan-s-glossary.md): Crack open the shell of predictive analytics with our nutty glossary, packed full of kernel knowledge for all your modeling needs!
- [Deleting Your Pecan Account](https://help.pecan.ai/en/articles/9268388-deleting-your-pecan-account.md): While we are sorry to hear that you've decided to remove or disable your account - we're here to make it smooth and simple
- [Inviting and Managing Team Members on Pecan](https://help.pecan.ai/en/articles/11534534-inviting-and-managing-team-members-on-pecan.md): Streamline collaboration by adding team members and controlling access permissions in your workspace
- [Data Sharing with Pecan: Your Complete Guide to Secure Predictive Analytics](https://help.pecan.ai/en/articles/11697668-data-sharing-with-pecan-your-complete-guide-to-secure-predictive-analytics.md): Everything you need to know about sharing data safely with our AI platform

## Connecting Your Data

- [Creating a data connection](https://help.pecan.ai/en/articles/6578109-creating-a-data-connection.md)
- [How to import data from your connection](https://help.pecan.ai/en/articles/6454476-how-to-import-data-from-your-connection.md): Feeding Pecan your data is a breeze! Select your tables, tailor your settings, opt for a reference column, and click 'Import data' ✨
- [Data importing configurations and methods](https://help.pecan.ai/en/articles/6800304-data-importing-configurations-and-methods.md): Import only new Rows, the full table, and overwrite - all to ensure accurate, efficient model training with easy method selection.
- [Deleting Files & Connections in Pecan](https://help.pecan.ai/en/articles/9251150-deleting-files-connections-in-pecan.md): Keep your Pecan environment tidy with our simple guide on removing unnecessary files and connections
- [What is a reference column, and when do you need one?](https://help.pecan.ai/en/articles/6454377-what-is-a-reference-column-and-when-do-you-need-one.md): Pecan optimizes table imports using reference columns to identify new rows, saving you time and resources.
- [Maintaining Data Integrity in Pecan Amidst Architecture Changes](https://help.pecan.ai/en/articles/6454479-maintaining-data-integrity-in-pecan-amidst-architecture-changes.md): Keep your model on track by aligning data changes with imported tables to avoid mismatches.
- [Data partitioning in Pecan](https://help.pecan.ai/en/articles/6691359-data-partitioning-in-pecan.md): Efficiently manage large datasets in Pecan with strategic data partitioning, enhancing query performance and prediction speed.
- [Enhanced Table Previews](https://help.pecan.ai/en/articles/8541578-enhanced-table-previews.md): Learn how to use Pecan's Enhanced Table Preview to get a quick, graphical summary and key stats of your imported data.
- [Uploading CSV files to Pecan](https://help.pecan.ai/en/articles/6827351-uploading-csv-files-to-pecan.md): Unleash the power of your CSVs with Pecan: easily upload, format, and get predictive insights. Perfect for quick modeling!
- [Converting files into CSV format](https://help.pecan.ai/en/articles/6827372-converting-files-into-csv-format.md)
- [AWS S3 Delta Lake tables](https://help.pecan.ai/en/articles/11459159-aws-s3-delta-lake-tables.md): Connect compressed, schema-flexible Delta Lake tables on AWS S3 to Pecan for faster, storage-smart analytics.
- [AWS S3 Parquet files](https://help.pecan.ai/en/articles/8181836-aws-s3-parquet-files.md)
- [Oracle connector](https://help.pecan.ai/en/articles/6656136-oracle-connector.md)
- [Snowflake connector](https://help.pecan.ai/en/articles/6454444-snowflake-connector.md): Learn how to securely connect Pecan to Snowflake using key-pair or username/password authentication methods.
- [Microsoft SQL Server connector](https://help.pecan.ai/en/articles/6454447-microsoft-sql-server-connector.md): Easily connect to your Microsoft SQL Server with Pecan, so you can create models using your existing data.
- [Google BigQuery connector](https://help.pecan.ai/en/articles/6454454-google-bigquery-connector.md)
- [IBM Db2 connector](https://help.pecan.ai/en/articles/6454456-ibm-db2-connector.md): Easily connect to your IBM Db2 with Pecan, so you can create models using your existing data.
- [Amazon Redshift connector](https://help.pecan.ai/en/articles/6454458-amazon-redshift-connector.md): Easily connect to your Amazon Redshift data warehouse with Pecan, so you can create models using your existing data.
- [PostgreSQL connector](https://help.pecan.ai/en/articles/6454460-postgresql-connector.md)
- [MySQL connector](https://help.pecan.ai/en/articles/6531807-mysql-connector.md)
- [HubSpot Connector](https://help.pecan.ai/en/articles/8126634-hubspot-connector.md)
- [ClickHouse database connector](https://help.pecan.ai/en/articles/8536663-clickhouse-database-connector.md): Easily connect to your ClickHouse Server with Pecan, so you can create models using your existing data.
- [Create a "Read" Connection To Salesforce](https://help.pecan.ai/en/articles/9292498-create-a-read-connection-to-salesforce.md)
- [Create a "Write" Connection To Salesforce](https://help.pecan.ai/en/articles/9264900-create-a-write-connection-to-salesforce.md): Optimize your CRM with predictive data—learn to connect, configure, and export Pecan insights directly to Salesforce
- [Databricks connector](https://help.pecan.ai/en/articles/11689905-databricks-connector.md)

## Creating a Model

- [Build your first Predictive ML model](https://help.pecan.ai/en/articles/8665302-build-your-first-predictive-ml-model.md): Blaze through the process of creating an AMAZING predictive ML model using Pecan's unique features - for pros and beginners alike!
- [Replacing mock data with real data \(if you used mock data\)](https://help.pecan.ai/en/articles/8843305-replacing-mock-data-with-real-data-if-you-used-mock-data.md): Learn to replace mock data with actual data with ease while keeping the notebook structure intact for a great predictive model in minutes
- [How to add attributes to your model](https://help.pecan.ai/en/articles/8656723-how-to-add-attributes-to-your-model.md): Attribute tables are an easy way to enrich your records with relevant data that will be valuable for generating predictions.
- [Creating a model | FAQ](https://help.pecan.ai/en/articles/6522326-creating-a-model-faq.md)
- [Recommended data volume for machine learning model](https://help.pecan.ai/en/articles/8100008-recommended-data-volume-for-machine-learning-model.md)
- [Understanding Pecan's Agent Memory](https://help.pecan.ai/en/articles/15849957-understanding-pecan-s-agent-memory.md): Pecan's agent now remembers facts about your data across conversations, so you don't have to re-explain them. Automatic, and shared across your org.
- [Understanding Draft Models](https://help.pecan.ai/en/articles/15888773-understanding-draft-models.md): Train fast draft models in Pecan to catch data leakage, test attributes, and check model health before committing to a full production training run.
- [The one-label training challenge](https://help.pecan.ai/en/articles/8896087-the-one-label-training-challenge.md): Ensure balanced data sets in models - key to successful training and accurate predictions
- [Configuring Train/Test Data Splits in Pecan](https://help.pecan.ai/en/articles/11381693-configuring-train-test-data-splits-in-pecan.md): Guide to time-based data partitioning for accurate machine-learning models
- [Feature engineering and description tags](https://help.pecan.ai/en/articles/6454522-feature-engineering-and-description-tags.md)
- [Splitting training data into Train, Validation, and Test Sets](https://help.pecan.ai/en/articles/6454518-splitting-training-data-into-train-validation-and-test-sets.md)
- [Introduction to selecting an optimization metric](https://help.pecan.ai/en/articles/6824084-introduction-to-selecting-an-optimization-metric.md)
- [Optimization metrics for binary models](https://help.pecan.ai/en/articles/6852262-optimization-metrics-for-binary-models.md)

## Evaluating a Model

- [Exploring Predictions with SQL Queries](https://help.pecan.ai/en/articles/10751104-exploring-predictions-with-sql-queries.md): Query your data to uncover trends over time, monitor prediction volumes, and discover insights by country or campaign.
- [Dashboard Overview For Binary Models](https://help.pecan.ai/en/articles/7325374-dashboard-overview-for-binary-models.md): Pecan's dashboard offers accuracy insights & customization, empowering you to optimize model performance & surpass benchmarks.
- [Evaluation metrics in binary models](https://help.pecan.ai/en/articles/7325418-evaluation-metrics-in-binary-models.md): Binary classification models use confusion matrix to calculate precision & recall rates, determining model effectiveness & accuracy.
- [Understanding Probability Score](https://help.pecan.ai/en/articles/7159702-understanding-probability-score.md): Binary classification in ML: Models predict entity classes with probability scores, and thresholds customize predictions for business needs.
- [Dashboard Overview For Regression Models](https://help.pecan.ai/en/articles/8058053-dashboard-overview-for-regression-models.md): Explore your Pecan AI model's insights with a dashboard: track performance, compare predictions, analyze feature importance, and more!
- [Model performance metrics for regression models](https://help.pecan.ai/en/articles/6456388-model-performance-metrics-for-regression-models.md): Master regression model evaluation with Pecan's diverse metrics: MdAPE, MAPE, WMAPE, WMPE, R2, and more. Precision in every prediction!
- [Outliers Alert in Dashboards \(Regression Models\)](https://help.pecan.ai/en/articles/10338954-outliers-alert-in-dashboards-regression-models.md): Spot outliers in regression models, clip them, and improve predictions with Pecan’s Outliers Alert
- [Understanding Pecan’s Benchmarks](https://help.pecan.ai/en/articles/7338218-understanding-pecan-s-benchmarks.md): Benchmarks evaluate ML models by comparing them to rule-based models, to understand their performance and communicate value to stakeholders
- [What is data leakage and how can you prevent it?](https://help.pecan.ai/en/articles/6457630-what-is-data-leakage-and-how-can-you-prevent-it.md): Avoid data leakage in ML models: Use only pre-prediction data to prevent "future peeking" and ensure valid, accurate outcomes.
- [What is overfitting?](https://help.pecan.ai/en/articles/6457619-what-is-overfitting.md): Overfitting is when a model memorizes training data instead of learning patterns. Resolve it by reducing attributes and adding more data.
- [What is underfitting?](https://help.pecan.ai/en/articles/6457626-what-is-underfitting.md): Underfitting occurs when a model fails to capture the underlying patterns in data, leading to poor performance on both training and new data
- [How do you know if your model is good?](https://help.pecan.ai/en/articles/7338168-how-do-you-know-if-your-model-is-good.md): To determine your model's performance, we compare its lift to random guess and benchmark models. You can also run A/B tests and outcomes.
- [How to determine if your model is healthy?](https://help.pecan.ai/en/articles/8223517-how-to-determine-if-your-model-is-healthy.md): Think of Pecan's health checks as a model's doctor visit. From data diets to overfitting sniffles, we ensure your model is in tip-top shape!
- [Understanding threshold logic](https://help.pecan.ai/en/articles/7325433-understanding-threshold-logic.md): Thresholds in binary classification models balance precision & recall, affecting model performance. Adjust based on business needs & costs.
- [Understanding Column importance](https://help.pecan.ai/en/articles/7835868-understanding-column-importance.md): A peak into the "black box" of a model: Key to unlocking model insights and optimizing predictions by weighting features' impact on outcomes
- [What is Label stability?](https://help.pecan.ai/en/articles/8099590-what-is-label-stability.md): Tackle Label Drift in ML models with Pecan: Detect shifts, adapt training, and monitor for precision in ever-evolving data landscapes.
- [Maintaining Feature Balance in Machine Learning Models](https://help.pecan.ai/en/articles/8193615-maintaining-feature-balance-in-machine-learning-models.md): Feature or column importance in ML models gauges predictor significance.
- [Understanding Explainability & Prediction Details](https://help.pecan.ai/en/articles/7936923-understanding-explainability-prediction-details.md): Entity-level explainability is a great tool for understanding and interpreting ML models, improving them and even help finding errors.
- [SHAP values](https://help.pecan.ai/en/articles/6457613-shap-values.md): SHAP values quantify feature impact in ML models, revealing key drivers in predictions and aiding in data-driven decision-making.
- [Model performance metrics for binary models](https://help.pecan.ai/en/articles/6456385-model-performance-metrics-for-binary-models.md): Learn about binary model metrics: Base Rate, Precision, Detection, AUC, LogLoss guide accurate, balanced predictions for distinct classes.
- [Understanding Area Under the Curve \(AUC\)](https://help.pecan.ai/en/articles/6457601-understanding-area-under-the-curve-auc.md)

## Generating Predictions

- [Using Your Model To Schedule Automated Prediction Cycles](https://help.pecan.ai/en/articles/15167169-using-your-model-to-schedule-automated-prediction-cycles.md): Download or send predictions back to your data warehouse for intelligent, quick, and confident data-driven decisions.