PyBI — Business Intelligence in Python
Install name:
pybi-ray· command:pybi· import:import pybi
PyBI is an open-source, Power BI–style business intelligence studio that you install with pip and run on your own computer. Connect data, shape it with a visual query editor, build a star-schema model, write DAX-style measures, design interactive dashboards, share a live link on your network, and export to PNG, PowerPoint, PDF or a self-contained interactive HTML file.
pip install pybi-ray
pybi run # opens PyBI Studio at http://127.0.0.1:8050
pybi demo # opens with a ready-made 3-page sales dashboard
Created by Muhammad Rafay Shaikh, from Pakistan.
More screenshots
| Get data (214 connectors) | Navigator & preview |
|---|---|
| Query editor | Model view |
Features
| Area | What you get |
|---|---|
| Get data | 214 connectors: files (Excel, CSV, JSON, XML, Parquet, PDF tables, SAS/SPSS/Stata…), 50 databases (SQL Server, PostgreSQL, MySQL, Oracle, Snowflake, BigQuery, Databricks…), NoSQL, cloud storage (S3, Azure, GCS, SFTP), 70+ SaaS apps (Salesforce, HubSpot, Stripe, Shopify, Jira, Notion…), public data (World Bank, FRED, Eurostat…), REST/GraphQL/OData/Python — with preview, navigator and refresh |
| Query editor (Power Query style) | Applied steps with live preview and column quality bars: choose/remove columns, promote headers, change type, filter rows, replace values, fill up/down, remove duplicates/blanks, keep top rows, custom column, conditional column, split column, text formatting, date parts, round, group by, unpivot, merge queries, append queries, plus a JSON advanced editor |
| Model view | Drag-and-drop relationships, many-to-one with single or both cross-filter direction, active/inactive, autodetect, one-click date table |
| DAX-style formulas | Measures and calculated columns with VAR … RETURN, CALCULATE, FILTER, ALL, ALLEXCEPT, ALLSELECTED, REMOVEFILTERS, KEEPFILTERS, VALUES, SELECTEDVALUE, RELATED, RELATEDTABLE, LOOKUPVALUE, iterators (SUMX, AVERAGEX, RANKX, TOPN, CONCATENATEX), time intelligence (TOTALYTD, DATESYTD, SAMEPERIODLASTYEAR, DATEADD, PARALLELPERIOD, PREVIOUSMONTH, DATESBETWEEN, …), logic, math, text and date functions |
| Visuals (25) | Clustered / stacked / 100% column and bar, line, area, stacked area, line & column combo, waterfall, ribbon, scatter / bubble, pie, donut, treemap, funnel, gauge, card, multi-row card, KPI, table (with data bars), matrix (expand/collapse, subtotals), slicers (list, dropdown, tiles, range), text box, smart narrative |
| Interactivity | Cross-filtering by clicking any chart (Ctrl+click for multi-select), slicers (optionally synced across pages), visual / page / report filters (basic, advanced, range), tooltips, focus mode, export data behind a visual |
| Report design | Drag, resize, snap & alignment guides, pages, page sizes, 6 themes (including dark), per-visual formatting, undo/redo, copy/paste, keyboard nudging |
| Smart features | Ask your data (type "top 5 products by profit as bar"), Auto dashboard (builds a full insight page from your model), Smart narrative (auto-written summary) |
| Share | Live, read-only network link (/view) for anyone on your Wi-Fi/LAN that refreshes as you edit; editing stays on your machine |
| Export | PNG (page or all pages), PowerPoint .pptx, PDF, interactive HTML (one offline file with data, slicers and cross-filtering), .pybi project file, CSV/Excel of any table |
Connectors (214)
Open Get data to search the catalog. Each connector shows a form, a live preview, and (for databases, workbooks, folders, buckets) a Navigator to pick tables. Loaded sources remember how they were connected, so Refresh re-pulls fresh data while keeping your query steps. Passwords and tokens are stored only on your computer in ~/.pybi/credentials.json — never inside .pybi files.
Connectors marked † need an extra driver (pip install … — PyBI tells you the exact command).
File (23) — Text / CSV, Tab-separated (TSV), Fixed-width text, Excel workbook, Excel 97-2003 (.xls) †, Excel binary (.xlsb) †, OpenDocument spreadsheet †, JSON, JSON Lines / NDJSON, XML, HTML file, Parquet †, Feather / Arrow IPC †, ORC †, Avro †, SAS (.sas7bdat / .xpt), SPSS (.sav) †, Stata (.dta), HDF5 †, YAML, PDF tables, ZIP archive, Folder (combine files)
Database (50) — SQLite database, PostgreSQL †, TimescaleDB †, Greenplum †, YugabyteDB †, Supabase (Postgres) †, Neon (serverless Postgres) †, Google Cloud SQL for PostgreSQL †, Amazon Aurora / RDS PostgreSQL †, Azure Database for PostgreSQL †, Google AlloyDB †, MySQL †, MariaDB †, PlanetScale †, TiDB †, Google Cloud SQL for MySQL †, Amazon Aurora / RDS MySQL †, Azure Database for MySQL †, Microsoft SQL Server †, Azure SQL Database †, Azure Synapse Analytics (SQL) †, Microsoft Fabric Warehouse (SQL endpoint) †, Oracle Database †, IBM Db2 †, SAP HANA †, Teradata †, Vertica †, CockroachDB †, SingleStore †, Exasol †, Firebird †, ClickHouse †, Trino †, Presto †, Apache Hive †, Spark SQL (Thrift server) †, Apache Impala †, Apache Druid †, Apache Pinot †, Dremio †, Snowflake †, Amazon Redshift †, Google BigQuery †, Amazon Athena †, Databricks SQL †, DuckDB †, MotherDuck †, Microsoft Access †, ODBC (any driver) †, Other SQL database (SQLAlchemy URL) †
Web & online files (10) — Web page tables, Wikipedia tables, File from URL, Google Sheets (shared link), Google Drive file, OneDrive / SharePoint file link, Dropbox file, GitHub file, Hugging Face dataset file, RSS feed
Public data (20) — Our World in Data, World Bank indicators, FRED economic data, CoinGecko crypto markets, CoinGecko price history, Open-Meteo weather history, REST Countries, Exchange rates (ECB), Alpha Vantage stocks, Yahoo Finance †, US Census Bureau, Socrata open data (SODA), CKAN open data (data.gov…), Eurostat, OECD data (SDMX), IMF DataMapper, USGS earthquakes, NASA POWER climate, Kaggle dataset †, Hacker News top stories
NoSQL & search (16) — MongoDB †, Amazon DocumentDB †, Elasticsearch, OpenSearch, Apache Solr, Apache CouchDB, InfluxDB 2.x / Cloud, Prometheus, Redis †, Apache Cassandra / ScyllaDB †, Amazon DynamoDB †, Google Firestore †, Firebase Realtime Database, Azure Cosmos DB (NoSQL) †, Neo4j (Cypher), Supabase (REST API)
Cloud storage (13) — Amazon S3 †, MinIO †, Cloudflare R2 †, Wasabi †, DigitalOcean Spaces †, Backblaze B2 (S3 API) †, Google Cloud Storage †, Azure Blob Storage †, Azure Data Lake Storage Gen2 †, SFTP †, FTP / FTPS, WebDAV / Nextcloud †, Hadoop HDFS †
Generic & developer (4) — REST API (any JSON / CSV API), GraphQL API, OData feed, Python script
Developer tools (8) — GitHub, GitLab, Bitbucket Cloud, Azure DevOps, Linear, Sentry, Datadog metrics, Grafana (dashboards list)
Project & productivity (12) — Jira Cloud, Confluence Cloud, Trello, Asana, ClickUp, monday.com, Notion database, Airtable, Smartsheet, Coda, Baserow, Slack
Support & ITSM (9) — PagerDuty, Zendesk Support, Freshdesk, Freshservice, Intercom, ServiceNow, Help Scout, Okta users, Twilio messages
CRM & sales (9) — Salesforce (SOQL), HubSpot, Pipedrive, Zoho CRM, Freshsales, Close CRM, Copper CRM, Dynamics 365 / Dataverse, ActiveCampaign
Finance & payments (8) — Stripe, PayPal transactions, Square, Xero, QuickBooks Online, Chargebee, Paddle Billing, Lemon Squeezy
Marketing & analytics (12) — Google Analytics 4, YouTube channel videos, Meta (Facebook) Ads insights, Mailchimp, Klaviyo, SendGrid stats, Brevo (Sendinblue), Typeform responses, SurveyMonkey, Jotform, Eventbrite, Calendly
E-commerce & CMS (8) — Webflow CMS, Shopify, WooCommerce, BigCommerce, WordPress, Contentful, Strapi, Gumroad sales
HR & time tracking (5) — BambooHR, Greenhouse, Harvest, Toggl Track, Clockify
Microsoft & Google (7) — Microsoft Graph (any endpoint), SharePoint Online list, Dynamics 365 Business Central, Microsoft Planner tasks, Google Sheets (API), Google Calendar, Google Search Console
Missing something? Four universal connectors reach almost anything: REST API (auth + 6 pagination styles + nested JSON flattening), GraphQL, ODBC / SQLAlchemy URL, and Python script.
SaaS connectors use the vendor's public REST API with a token you provide (PyBI does not run OAuth sign-in flows yet). They were built from each vendor's API documentation and tested against a mock API, not against live accounts — if a vendor changes an endpoint, override it under Advanced options (endpoint, records path, query parameters) and please open an issue.
Install drivers in bulk with extras, e.g. pip install "pybi-ray[postgres,mysql,cloud]" or pip install "pybi-ray[all]".
Usage
pybi run # start the studio (network view link enabled)
pybi run sales.xlsx # start with a data file loaded
pybi run my_report.pybi # open a saved project
pybi run --private # only this computer can connect
pybi run --port 9000 --no-browser
pybi demo # demo dataset + dashboard
pybi export my_report.pybi -o dashboard.html # headless export to interactive HTML
python -m pybi run # same as `pybi run`
From Python or Jupyter:
import pandas as pd
import pybi
sales = pd.read_csv("sales.csv")
products = pd.read_excel("products.xlsx")
pybi.run(Sales=sales, Products=products) # opens the studio with these tables
Example measures
Total Revenue = SUM(Sales[Revenue])
Profit Margin = DIVIDE([Profit], [Total Revenue])
Revenue LY = CALCULATE([Total Revenue], SAMEPERIODLASTYEAR('Calendar'[Date]))
Revenue YTD = TOTALYTD([Total Revenue], 'Calendar'[Date])
% of Total = DIVIDE([Total Revenue], CALCULATE([Total Revenue], ALLSELECTED(Sales)))
Product Rank = RANKX(ALL(Products), [Total Revenue])
Big Orders = CALCULATE([Orders], Sales[Quantity] >= 3)
Calculated column:
Sales[Cost] = Sales[Quantity] * RELATED(Products[UnitCost])
How it works
- Python (Flask + pandas) serves the studio, loads and transforms data (query steps are replayable and stored in the project), saves
.pybiprojects (a zip of raw data, steps and the report definition), and builds PPTX/PDF/HTML exports. - The analytics engine (filter context, relationships, the DAX-style language) and the d3-based visuals run in the browser, so the studio, the network view and exported HTML files behave identically.
- An autosave of the current project is kept in
~/.pybi/autosave.pybi.
Security notes
- The studio binds to all network interfaces so the
/viewlink works on your LAN. Requests from other computers can only view (/view, read-only data); all editing APIs are restricted to the local machine. Use--privateto disable network access completely. - The custom-column step in the query editor evaluates pandas expressions. Only open
.pybifiles from people you trust.
Limitations (v0.1)
PyBI is not a Microsoft product and does not open .pbix files. Data is held in memory, which works well up to a few million rows depending on your machine. Not every DAX function is implemented; unsupported functions show a clear error in the formula editor.
Contributing
PyBI is built in the open and contributions are very welcome — connectors tested with real accounts, visuals, DAX functions, docs and translations.
- Read CONTRIBUTING.md for setup, project structure and how to add connectors/visuals.
- Look for issues labelled good first issue.
- Questions and ideas: Discussions.
git clone https://github.com/rafay183/pybi
cd pybi
pip install -e ".[dev]"
pytest -q
pybi demo
Maintainers: release steps are in docs/PUBLISHING.md. See also CHANGELOG.md and SECURITY.md.
License
MIT © Muhammad Rafay Shaikh. Bundles d3 (ISC license, © Mike Bostock).
Metadata
Release files for pybi-ray 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| pybi_ray-0.1.0.tar.gz | 258.1 kB | With provenance | Details for pybi_ray-0.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| pybi_ray-0.1.0-py3-none-any.whl | Python 3 | none | any | Details for pybi_ray-0.1.0-py3-none-any.whl |
Total release size: 510.1 kB
Release files / pybi_ray-0.1.0.tar.gz
| Download URL | pybi_ray-0.1.0.tar.gz |
|---|---|
| Size | 258.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
5bb9de3c5084bb823b74a2897bf910e02db8cfc33643d810b4135220bab26af3
|
|
BLAKE2b-256 checksum How to use checksums |
7d588f817b5ca98396ad946c1d813012e230ee39264cb16ab9eec8284430fd61
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 12, 2026.
Transparency logRelease files / pybi_ray-0.1.0-py3-none-any.whl
| Download URL | pybi_ray-0.1.0-py3-none-any.whl |
|---|---|
| Size | 252.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
17cde991aa7dc8f49f76ec129ec778a73b33ddaa59ad936929576d3eb426f848
|
|
BLAKE2b-256 checksum How to use checksums |
b95063cb2c42a9785f8056869a5cb81bd5d0523b946ac4cd7f726a619976c01d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 12, 2026.
Transparency logSecurity for pybi-ray 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
Both files in this release have been uploaded with provenance metadata.
Attestation for pybi_ray-0.1.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 12, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.