Oxygen is the Next-Generation Data Platform.
We're building the Postmodern Data Stack: Agent-Native, Full-Stack, Composable, Lightweight, and Open (AFCLO).
What do you want to know about your data?
Try asking
Used by Fortune 500 companies and high-growth startups
The Postmodern Data Stack. AFCLO.
Five principles that define the next generation of data infrastructure.
Agent-Native
Automate all parts of the data analytics lifecycle with purpose-built AI, integrated natively with the stack. Do work 100 to 1000 times faster than before.
Full-Stack
Say goodbye to the 5-piece Modern Data Stack. BI, semantic layer, data transformation, data ingestion, and data warehousing — all in one integrated platform.
Composable
Modular by design. Mix and match with your existing tools like Snowflake, Fivetran, dbt, Looker, and Hex.
Lightweight
Oxygen runs light, powered by Rust and DuckDB. Say goodbye to clunky Spark clusters and the 6 to 7-figure Snowflake bill. Fast, efficient, and doesn't break the bank.
Open
Open-source and built on open standards. Powered by Rust, DuckDB, Postgres, and Parquet. No vendor lock-in, full transparency.
Intelligence powered by the Large Analytics Model.
Oxygen achieves deterministic precision through the Large Analytics Model (LAM) — a purpose-built system for data analytics based on the principle of entropy reduction.
Semantic and procedural context grounding every query in your business truth
Verifiable logic chains for traceable, deterministic results
Precise translation from natural language to analytics operations
Context
Semantic and procedural grounding
Reasoning
Verifiable logic chains
Inference
Precise NL translation
Full-Stack and Composable. All-In-One or Mix & Match.
Oxygen is a complete data platform with BI, semantic layer, transformation, storage, and ingestion built-in. Use it as an integrated stack, or compose it with your existing tools — adopt incrementally, replace when ready.
Semantic Layer & BI
Looker
Cube
Omni
Hex
Airlayer
AirappTransformation
dbt
AirformData Ingestion
Fivetran
Airbyte
dlt
AirwayData Warehouse
Snowflake
BigQuery
ClickHouse
Databricks
AirhousePlug in your existing tools — mix and match with Oxygen components
Data Apps
AirappSemantic Layer
AirlayerTransformation
AirformData Ingestion
AirwayData Warehouse
AirhouseZero-config, fully integrated — everything you need in one platform
Lightweight by design. Powered by Rust and DuckDB.
Forget clunky distributed computing clusters. Oxygen delivers the power of the Modern Data Stack at a fraction of the cost and complexity.
Single Binary
The entire platform compiles to a single binary. No JVM, no Python dependencies, no cluster coordination.
In-Process Execution
Query execution happens in-process with DuckDB. No network hops, no serialization overhead, no cold starts.
Blazing Fast
Rust's zero-cost abstractions and DuckDB's vectorized engine deliver performance that rivals distributed systems.
10x Cost Savings
No Spark clusters, no Kubernetes overhead, no cloud compute sprawl. Run on a single machine what used to require a fleet.
Modern Data Stack
7+ services, distributed infrastructure, 6-7 figure annual cost
Oxygen Stack
One process, in-memory execution, 10x cost savings
Open Source and Open Format. No vendor lock-in.
Built on proven open-source technologies and open standards. Your data stays portable, your stack stays flexible.
Rust
Blazing fast, memory-safe core engine built for performance and reliability.
Parquet
Open columnar storage format for efficient analytics and cross-tool compatibility.
DuckDB
Embedded analytical database for lightning-fast in-process queries.
Postgres
Battle-tested relational database for metadata and application state.
Flexible Deployment Options
Run fully managed in the cloud or self-host the complete open-source stack.
Oxygen Cloud
Fully managed cloud service. No infrastructure to manage, automatic updates, and enterprise-grade security.
Oxygen OSS
Self-deploy the complete stack including all engines, semantic layer, BI, and AI capabilities.
View on GitHubSee Oxygen in Action
Explore how teams use Oxygen to transform their data workflows.
Automate the entire data lifecycle. From raw data to business value.
Oxygen handles everything end-to-end — data integration, transformation, semantic modeling, and delivering data products. One platform, zero handoffs.
Extract & Load
AlphaConnect to any source and ingest data into your data warehouse or lakehouse.
Transform
AlphaClean, normalize, and prepare data with AI-assisted transformations.
Semantic Model
Define business logic, metrics, and relationships in a unified layer.
Data Products
Ship dashboards, apps, automations, and agent-powered workflows.
Show me the factors influencing revenue growth
To identify the factors influencing revenue growth, it's important to analyze how various conditions—such as economic indicators (fuel prices, unemployment rate, CPI), seasonality (month, holiday weeks), and environmental factors (temperature)—correlate with and impact total sales over time.
Step-by-step plan:
- Check correlations between sales and major external factors: temperature, fuel prices, unemployment, and CPI.
- Compare sales during holiday weeks vs. non-holiday weeks.
- Observe sales trends across months (seasonality).
- Summarize the impact of these factors (correlations, seasonal patterns, holiday effects).
I'll query and synthesize results for:
- Correlation coefficients for each economic/environmental factor.
- Sales uplift during holidays vs. regular weeks.
- Revenue distribution by month.
Executing the relevant queries in parallel.
query_retail_analytics
Click to view the artifact.
query_retail_analytics
Click to view the artifact.
query_retail_analytics
Click to view the artifact.
Here are the key factors influencing revenue growth, based on available data:
- Sales show notable variation by month, with peak revenues in July, April, and March. The lowest sales are typically in January and November.
- This suggests strong seasonal effects, possibly tied to holidays, back-to-school, and summer promotions.
- The average revenue during holiday weeks is higher ($1,122,888) than during non-holiday weeks ($1,041,256).
- There is a clear holiday sales lift: sales are 7.84% higher during holiday weeks versus non-holiday weeks.
- The correlation between temperature and sales is slightly negative (-0.064). This means hotter or colder weeks have a minor inverse relationship with sales, but it's not strong.
- The correlation between fuel prices and sales is nearly zero (0.009), suggesting very little direct impact.
- There is a weak negative correlation (-0.106) between unemployment rate and sales, implying that higher unemployment might slightly reduce sales, but it is not a dominant factor.
- The average values for context: temperature ~61°F, fuel price ~$3.36/gallon, unemployment ~8.0%, CPI ~171.58.
Summary:
- The biggest drivers of revenue growth are seasonality (time of year) and holidays.
- Economic factors like unemployment do have a small effect, while temperature and fuel prices have minimal influence on revenue in this dataset.
- Understanding these factors can inform marketing, promotions, and inventory strategies.
Would you like to see this visually (e.g., revenue by month chart or more detail on any specific factor)?
Ask any data question. Get precise, reasoned answers.
Oxygen agents analyze your data with the rigor of a senior analyst. They build step-by-step plans, execute parallel queries, and synthesize findings into actionable insights — all in seconds.
Automate analytics workflows with agentic procedures
Define complex multi-step analytics workflows as code. Oxygen procedures orchestrate SQL queries, transformations, and AI agents to produce deterministic, reproducible results every time.
Revenue Performance Analysis Report
Generated:
Executive Summary
Store Network Overview
- Total Stores: 45
- Status: Active
Revenue Performance
- Total Revenue: $673721898.11
- Average Revenue per Store: $14971597.49
- Average Weekly Revenue: $1046964.88
- Analysis Period: 14645 to 15639
Top Performing Stores
#1: Store 20
Build anything from ETL pipelines to data apps
Use natural language to build ETL pipelines, semantic models, interactive dashboards, and custom data applications. Oxygen generates production-ready code and deploys instantly.
Build me a data app looking at the correlations between factors that affect revenue and revenue
I built a data app to analyze the correlations between various factors (Temperature, Fuel Price, CPI, Unemployment, and Holidays) and store revenue, measured as Weekly Sales.
Here's what the app does:
- Shows a summary table of correlation coefficients between Weekly Sales and each factor.
- Plots average weekly sales against each factor (Temperature, Fuel Price, CPI, Unemployment) to visualize trends.
- Compares average revenue on holidays vs. non-holidays using a bar chart.
These visualizations and statistics help to quickly understand how much each factor correlates with revenue and reveal any potential patterns or relationships in the data.
I regenerated the data app revenue_correlations_app.app.yml with the following features and plots:
- A summary table showing the correlation coefficients between Weekly Sales and each factor (Temperature, Fuel Price, CPI, Unemployment, Holiday Flag).
- Bar chart comparing average weekly sales on holidays vs. non-holidays.
- Four line charts showing average weekly sales as a function of each factor:
- Temperature
- Fuel Price
- CPI
- Unemployment
This setup allows you to clearly see statistical relationships and trends between revenue and the influencing variables, providing both quantitative (correlation table) and visual insights (charts).
Revenue Correlation Analysis
This dashboard visualizes how various factors relate to store revenue (Weekly Sales) at Oxymart. Use the plots and summary table to explore the relationships.
Correlation Coefficients with Weekly Sales
| corr_temp | corr_fuel | corr_cpi | corr_unemployment | corr_holiday |
|---|---|---|---|---|
| -0.06 | 0.01 | -0.07 | -0.11 | 0.04 |
Average Weekly Sales: Holiday vs. Non-Holiday
Average Weekly Sales by Temperature
Average Weekly Sales by Fuel Price
One platform for the whole org. AI agents for every team.
Oxygen agents work across your entire organization—empowering data teams to move faster and business teams to self-serve insights without waiting on tickets.
Data Teams
Speed up work by running data integration, transformation, semantic modeling, and data productization.
Product Teams
Self-answer questions and self-build data apps about product analytics.
Revenue Teams
Self-answer questions and self-build data apps about revenue analytics.
Marketing Teams
Self-answer questions and self-build data apps about marketing analytics.
Finance Teams
Self-answer questions and self-build data apps with live financial and operational data.
Every Team
No more bottlenecks. Everyone gets answers faster.
Get started today
Meet Your New Data Stack
Oxygen replaces 5 tools and consolidates them into one agent-native, lightweight stack. BI, semantic layer, transformation, storage, compute, and ingestion — all in an integrated platform.