INTELLIGENCE
Archive
How to Get Started with Mixpanel for Product Analytics
Mixpanel is an event-based product analytics platform that tracks user actions — signups, feature clicks, purchases — and turns them into funnels, retention charts, and cohort reports. Getting started requires three steps: mapping your event taxonomy before touching code, installing the SDK or connecting via Segment, and building your first three core reports. Teams that follow this sequence typically surface their first actionable retention insight within five to seven days of setup.
How to Analyze Graph Data with BigQuery Graph
BigQuery Graph, released in preview in April 2026, lets you model and query relationships in your existing BigQuery data using Graph Query Language (GQL) without moving or copying any data. You define nodes and edges as logical views over your tables, then write GQL pattern-matching queries directly in BigQuery Studio. This guide walks through creating a property graph, writing your first GQL query, and visualizing results for common use cases like fraud detection and supply chain tracing.
How to Use the =AI() Function in Google Sheets
Google Sheets includes a native =AI() function, powered by Gemini, that lets you run text generation, summarization, and categorization directly inside spreadsheet cells. You write a plain-English instruction in the formula, reference a cell range, and Gemini processes each row automatically. The function requires a Google Workspace Business Standard plan or higher and is currently available in the US only.
How to Get Started with Copilot in Power BI
Copilot in Power BI lets business users ask questions about their reports in plain English, generate new report pages from a description, and get instant summaries without writing a single formula. To use it, your organization needs a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher). This guide covers what Copilot can do, how to confirm your setup, and how to get useful results from day one.
How to Set Up ClickHouse Cloud for Analytics
ClickHouse Cloud is a managed deployment of the open-source ClickHouse columnar database, built for fast analytical queries on large datasets. You can create a free account, upload a CSV file, and run sub-second SQL aggregations without managing any infrastructure. This guide covers account setup, table creation, data loading from CSV and remote sources, running aggregation queries, and connecting a BI tool to your ClickHouse service.
How to Query S3 Files with Amazon Athena
Amazon Athena lets you run SQL directly on files stored in S3 with no database to set up or maintain. You point Athena at a bucket, define a table schema in the AWS Glue Data Catalog, and start querying CSV, JSON, or Parquet files within minutes. Athena charges $5 per terabyte scanned, so converting raw CSV to Parquet before querying can cut costs by 60 to 80 percent on the same workload.
How to Set Up Data Quality Checks with Soda
Soda is an open-source data quality platform that scans databases, warehouses, and CSV files using a YAML-based language called SodaCL. Install the library with a connector for your data source, write a checks file defining row count, freshness, and validity rules, then run soda scan from your terminal. As of February 2026, the library ships an AI agent called soda ai that generates checks automatically from your table schema or dbt models.
How to Get Started with Polars for Data Analysis
Polars is a Python DataFrame library built in Rust that runs queries in parallel across all CPU cores, making it 5 to 50 times faster than pandas on datasets above a few hundred thousand rows. Install with pip install polars, load a CSV with pl.read_csv(), and write expressions using .select(), .filter(), and .group_by(). Unlike pandas, Polars has no row index and supports lazy evaluation via .lazy() and .collect() for large file processing.
How to Set Up Cube for Agentic Analytics
Cube is an open-source semantic layer that sits between your database and your BI tools, defining metrics once so every query and AI agent uses the same numbers. You can run Cube locally with Docker in under ten minutes, connect it to a Postgres or warehouse database, generate a data model automatically, and query it in plain English using Cube's D3 agentic analytics layer. This guide walks through the complete setup from scratch.