EDA

Open-source business intelligence, honestly reviewed. No marketing fluff, just what you get when you self-host it.

TL;DR

• What it is: Open-source (AGPL-3.0) metadata-driven analytics and dashboarding platform — think Metabase or Redash, but aimed at making data exploration genuinely accessible to non-technical users [README].
• Who it’s for: Small business teams, operations managers, and founders who need to turn a production database into readable dashboards without writing SQL. Also useful for developers who want to expose data to stakeholders without building a custom reporting layer [2].
• Cost savings: Paid BI tools (Tableau, Looker, Power BI) run $15–70/user/month at team scale. EDA self-hosted is free software on a $5–10/mo VPS — data not available on their current SaaS pricing tiers.
• Key strength: Automatic data model generation — EDA reads your database schema and proposes a business model, so non-technical users never have to understand joins [README][2].
• Key weakness: Small community (188 GitHub stars), AGPL-3.0 license restricts commercial embedding, MongoDB dependency adds infrastructure overhead, and limited third-party documentation makes troubleshooting harder than with more established alternatives.

---

What is EDA

EDA — now officially called Edalitics, rebranded from “Enterprise Data Analytics” — is a web-based analytics platform built by Jortilles, a Barcelona-based open-source analytics company. The project’s stated purpose is blunt: “to be the simplest analytics system to use, even for users without technical skills” [README].

The core idea is metadata-driven analysis. You point EDA at a database (Postgres, MySQL, MariaDB, BigQuery, Oracle, SQL Server, Vertica), and it reads your schema to automatically generate a business-friendly data model — tables become “entities,” columns become “attributes,” and relationships are inferred. A non-technical user then navigates this model visually, picks what they want to see, and EDA builds the query behind the scenes [2][README].

The technology stack is Node.js/TypeScript backend, Angular frontend, and MongoDB for storing metadata and configuration. That last choice is worth noting: MongoDB is where EDA stores its own internal state (data models, dashboard configs, user definitions), not where your business data lives. Your actual data stays in whichever database you connected [README].

As of this review, the project sits at 188 GitHub stars and is licensed under AGPL-3.0. There is a managed SaaS option at free.edalitics.com for those who want to try before self-hosting [README].

---

Why people choose it

There is only one substantial third-party review of EDA/Edalitics available for synthesis. The opensource.com write-up [2] — authored by Juanjo Ortilles, one of Jortilles’ own developers — lays out the founding philosophy: the team spent years in business analytics and kept hitting the same problem. “The modern world collects data compulsively but there was no easy way for average people to see or interpret that data. There are some powerful open source tools for this purpose, but they are very complex.”

That framing puts EDA in a specific lane. It is not competing with Grafana (infrastructure metrics, time-series), not competing with Superset (power users, SQL-first), and not competing with Redash (developer-oriented, query-centric). The explicit target is a business user who already has data in a database and needs a clean window into it without calling a developer.

The feature that most distinguishes it in the opensource.com review is the automatic model generation [2]: connect to a database, and EDA proposes the model for you. This is something Metabase also does (and more visibly), but EDA’s pitch is that the model more closely mirrors how the business already thinks about the data. If your database is well-designed, EDA’s auto-generated model is usable without further configuration [2][README].

Row-Level Security is another differentiator cited in the README. The ability to define access down to the individual row means a single EDA instance can serve multiple departments or clients where each sees only their slice of the data [README].

---

Features

Based on the README and the opensource.com review:

Core analytics:

• No-code dashboard builder with drag-and-drop panels [README]
• Automatic data model generation from existing database schema [README][2]
• Tree-mode visual explorer for navigating the logical model [README]
• SQL mode for advanced users who want to bypass the no-code layer [README][2]
• Standard chart types: bar, pie, line, treemap [2]
• Map integration with geoJSON shapefiles, latitude/longitude support [2]

Alerts and sharing:

• KPI definitions with automatic email alerts when thresholds are crossed [README][2]
• Public dashboards shareable via URL link — no login required for the viewer [README][2]
• Private vs. public dashboard controls [2]

Data access and security:

• Row-Level Security (RLS) to grant users access to specific data rows [README][2]
• Data caches and programmatic refreshes to avoid hammering production databases [2]

Database support:

• Postgres, MySQL, MariaDB, SQL Server, Oracle, Vertica, BigQuery [README][2]
• MongoDB is used only for EDA’s own metadata storage, not as a queryable data source [README]

Technical stack:

• Node.js, TypeScript, Angular [README]
• REST API [merged profile]
• Responsive UI [merged profile]
• Docker, Helm deployment [README][merged profile]

What’s missing (compared to established alternatives):

• No native git-based version control for dashboards
• No embedded analytics SDK mentioned
• No SSO/SAML mentioned in available documentation
• Plugin or extension framework not documented

---

Pricing: SaaS vs self-hosted math

EDA/Edalitics SaaS:

• Free tier available at free.edalitics.com [README]
• Paid SaaS at edalitics.com — specific tier pricing not publicly listed in available sources (data not available at time of writing)

Self-hosted (AGPL-3.0):

• Software: $0
• A VPS to run it: $5–15/month (needs to host both the Node.js app and a MongoDB instance)
• Your time for setup and maintenance

Alternatives for comparison:

• Metabase Cloud: $500/month for 5 users on Pro; open-source self-hosted is free (SSPL license)
• Redash: free self-hosted, cloud tier pricing discontinued (acquired by Databricks)
• Apache Superset: free self-hosted (Apache 2.0); managed offerings from third parties
• Tableau Cloud: $70/user/month (Creator), minimum 1 user
• Looker Studio: free (Google, with limitations), Looker proper is enterprise-priced

The honest savings calculation for EDA self-hosted depends entirely on what you’re replacing. If you’re currently paying $200–500/month for Tableau or Power BI licenses for a 5-person team, a $10/month VPS running EDA is a significant reduction — assuming EDA covers your use cases. If you’re on Metabase free self-hosted already, the switch to EDA buys you little on cost and requires a migration.

The AGPL-3.0 license is worth understanding before committing: if you embed EDA in a SaaS product you distribute to customers, your entire application must also be AGPL-licensed or you need a commercial agreement with Jortilles. This is different from MIT or Apache 2.0. For internal use only, AGPL poses no issue [README].

---

Deployment reality check

The README offers a single-command Docker path: docker run -p 80:80 jortilles/eda:latest. That is genuinely as simple as self-hosted BI tools get for initial testing [README].

Production deployment is more involved:

What you actually need:

• A Linux VPS (2GB RAM minimum; 4GB if running MongoDB alongside the app)
• Docker installed
• A separate MongoDB instance — either in the same docker-compose or externally managed
• A domain and reverse proxy (nginx or Caddy) for HTTPS
• An SMTP setup if you want email alerts

One-click options exist via DeployStack for AWS, DigitalOcean, and Render, plus a Helm chart for Kubernetes [README]. These lower the barrier significantly for teams already on one of those platforms.

What can go sideways:

• MongoDB is an additional service to run, monitor, and back up. If you’re already running Postgres for your business data, adding MongoDB for EDA metadata is real infrastructure overhead.
• The database configuration is hardcoded in a file (EDA/eda/eda_api/config/database.config.js) rather than environment variables in the default setup — not cloud-native defaults [README]. This matters if you’re doing containerized deploys where you expect secrets to come from environment variables.
• Backend URL is configured in the Angular frontend source file, not an env variable [README]. Meaning if you change the backend URL after building, you need to rebuild the frontend.
• Documentation is available at doc.edalitics.com, but the project is small enough that community support (Stack Overflow threads, Reddit posts, GitHub issues) is sparse compared to Metabase or Superset.

Realistic time estimate: a technical user should have it running in 30–60 minutes via Docker. For a non-technical founder, budget 2–4 hours including DNS setup, or have someone deploy it once.

---

Pros and Cons

Pros

• Auto-generates the data model. EDA reads your database schema and proposes a business model automatically. For teams with a well-designed database, this means you can go from “connected to database” to “non-technical user browsing data” in under an hour [README][2].
• No-code first, SQL escape hatch. The drag-and-drop interface handles most cases; SQL mode is there when the model isn’t enough [README][2]. Both users coexist on the same instance.
• Row-Level Security. Restricting data access to specific rows per user is built in — this matters when one instance needs to serve multiple teams or clients who shouldn’t see each other’s data [README].
• Public shareable dashboards. Sharing a link to a public dashboard without requiring the viewer to log in is simple and useful for external stakeholders [README][2].
• Email KPI alerts. Define a threshold on a metric and have EDA notify the relevant person automatically — no third-party integration required [README][2].
• One-command Docker install. docker run -p 80:80 jortilles/eda:latest genuinely works for testing [README].
• Multiple database connections. Postgres, MySQL, BigQuery, Oracle, SQL Server, Vertica, MariaDB — the major options for business databases are covered [README][2].

Cons

• 188 GitHub stars. This is a small project. Established alternatives like Metabase (40K+ stars), Superset (63K+ stars), and Redash (26K+ stars) have vastly larger communities, more contributors, and more available help online. If EDA breaks in production, you’re mostly on your own.
• AGPL-3.0 is the most restrictive common open-source license. Any commercial redistribution or SaaS-embedding requires either open-sourcing your entire stack or buying a commercial license from Jortilles. MIT and Apache 2.0 projects don’t have this constraint [README].
• MongoDB dependency for internal state. Running a document store just to hold EDA’s own configuration is extra infrastructure. If your organization standardizes on Postgres, adding MongoDB is friction.
• Configuration not fully environment-variable-driven. The default setup requires editing source files for database URLs and backend URLs — a step backwards from modern 12-factor app expectations [README].
• Limited third-party reviews and community resources. Almost no independent user reviews, comparisons, or troubleshooting threads exist. Evaluating this tool based on community experience is difficult.
• No publicly listed SaaS pricing. The managed cloud product at edalitics.com does not show pricing on the public-facing page, making the buy-vs-self-host comparison harder to calculate (data not available at time of writing).
• Rebranding ambiguity. The GitHub repo is still jortilles/eda, the Docker image is jortilles/eda, the README title says “Edalitics,” the domain is edalitics.com, but search results surface old “Enterprise Data Analytics” branding. This creates confusion for anyone trying to research the tool.

---

Who should use this / who shouldn’t

Use EDA if:

• You have a well-structured relational database and a non-technical team that needs to explore it without developer involvement.
• You want a clean no-code dashboard interface and are willing to trade community size for simplicity.
• Your use is entirely internal — AGPL doesn’t create complications for internal tools.
• You’re already running MongoDB or don’t mind adding it to your stack.
• You want Row-Level Security across a single multi-tenant database without managing separate instances.

Skip it (pick Metabase instead) if:

• You want the same non-technical-friendly experience with 200x more community documentation, more integrations, and a more battle-tested product.
• You’re a team that might grow and need SSO, embedding, or enterprise features later.

Skip it (pick Apache Superset instead) if:

• Your team has technical users who want a SQL-first, charting-heavy tool with a large plugin ecosystem.
• You need a tool where you can find help on Stack Overflow and Reddit.

Skip it (pick Grafana instead) if:

• Your data is time-series metrics, logs, or infrastructure telemetry rather than business database records.

Skip it (pick Lightdash instead) if:

• You’re already using dbt and want a BI layer that treats dbt models as its source of truth.

---

Alternatives worth considering

• Metabase — the most direct comparison. Similar no-code philosophy, auto-generates models, far larger community, cleaner documentation. SSPL license for self-hosted (more restrictive than AGPL for commercial use). Commercial cloud pricing is high. For most teams, Metabase Community Edition is the default choice in this category.
• Apache Superset — SQL-first, powerful, Apache 2.0 licensed. Steeper learning curve than EDA or Metabase. Requires more devops effort. Strong for technical teams.
• Redash — developer-oriented, query-centric, Apache 2.0. Acquired by Databricks; cloud tier discontinued. Still actively maintained for self-hosted use. Good for teams comfortable writing SQL.
• Lightdash — MIT licensed, dbt-native. If you’re using dbt, this is a better fit than EDA. Requires existing dbt infrastructure.
• Grafana — not a direct comparison but often chosen by mistake when what’s needed is business BI rather than infrastructure monitoring. Excellent for time-series/metrics; awkward for relational business data.
• Tableau / Power BI — the incumbent commercial options. Better visualizations, enterprise support, and integrations, at $15–70/user/month. The cost pain is real at scale.

---

Bottom line

EDA/Edalitics has a genuine and useful idea: make a production database browsable by non-technical people, automatically, without writing SQL. The auto-model-generation and Row-Level Security are real features that take time to build in competing tools. For a small team that wants a lightweight internal analytics layer and is comfortable with Docker, it’s worth testing.

The honest caveat is size: 188 stars means limited community support, limited third-party documentation, and real risk if the project slows development. If self-hosting BI is the goal, Metabase Community Edition is the safer bet — larger community, more features, better documented. EDA makes sense if you’ve tried Metabase and found it too heavy, or if the auto-model approach specifically solves a workflow problem that simpler tools don’t. The AGPL license is fine for internal use but rules out any commercial embedding without a commercial agreement.

If the deployment overhead is the blocker, that’s exactly what unsubbed.co’s parent studio upready.dev handles for clients. One setup, you own the infrastructure.

---

Sources

1. Adam Kovac, Semiconductor Engineering — “Security Concerns Weigh Down Open-Source EDA” (October 10, 2024). https://semiengineering.com/security-concerns-weigh-down-open-source-eda/ — Note: this article covers Electronic Design Automation (chip design tools), a separate field from the Edalitics analytics platform reviewed here; not cited in the review body.
2. Juanjo Ortilles, Opensource.com — “Make your data boss-friendly with this open source tool” (April 2021). https://opensource.com/article/21/4/visualize-data-eda
3. FOSSi Foundation — “Roadmap and Recommendations for Open Source EDA in Europe” (November 2024). https://fossi-foundation.org/resources/eu-roadmap — Note: covers Electronic Design Automation, not the Edalitics analytics platform; not cited in the review body.
4. Helali et al., ACL Anthology / EMNLP 2025 — “Reliable and Cost-Effective Exploratory Data Analysis via Graph-Guided RAG”. https://aclanthology.org/2025.emnlp-main.836/ — Note: academic paper on automating data science EDA workflows, not about the Edalitics dashboarding tool; not cited in the review body.
5. Vikram Sekar, Vik’s Newsletter — “How to Learn Chip Design With Open-Source EDA Tools” (April 28, 2025). https://www.viksnewsletter.com/p/how-to-learn-chip-design-with-foss-eda — Note: covers Electronic Design Automation for semiconductor chip design; not cited in the review body.

Primary sources:

• GitHub repository and README: https://github.com/jortilles/eda (188 stars, AGPL-3.0)
• Free SaaS: https://free.edalitics.com
• Official website: https://www.edalitics.com
• Documentation: http://doc.edalitics.com