Project Details

Overview: Interactive Power BI dashboard giving a 360° view of key business metrics for a consumer goods company.
Views: Finance, Sales, Marketing, Supply Chain and Executive dashboards, plus Home, Support and Information pages for easy navigation and help.
Insights: P&L analysis, gross margin and net profit tracking, customer and product performance, segment profitability, and forecast accuracy with out-of-stock and excess-inventory risk indicators.
Build: Star-schema data model, Power Query for data loading and transformation, and DAX for KPIs and time intelligence.
Power BI 4.0 Upgrade: Data extended to FY2026 (about 3 million rows), incremental refresh so only new and recent data reloads, and Row-Level Security for role-based data access.
Outcome: Turns raw data into clear visual insights, helping finance, sales, marketing and leadership teams make faster, data-driven decisions.

Live Interactive Dashboard

Project Details

Overview: Exploratory data analysis of 134K+ bookings across 25 AtliQ Grands hotels in Delhi, Mumbai, Bangalore and Hyderabad, built to show where occupancy and revenue were won or lost after revenue went flat.
Tools: Python, Pandas, Matplotlib and Jupyter Notebook.
Data Preparation: Merged five datasets, removed invalid guest counts and extreme revenue outliers, filled missing capacity with the median, standardised mixed-format dates and appended new August data.
Analysis: Occupancy by city, room category and weekday vs weekend; revenue by city, month, property and booking platform; average ratings by city.
Key Findings: Weekend occupancy (72.3%) far exceeds weekday (50.9%); Mumbai leads total revenue (about ₹669M); Delhi has the top occupancy and rating; Bangalore is the lowest rated.
Recommendations: Weekend premium pricing, weekday promotions, direct booking growth, and a service-quality review in low-rated markets.

Project Details

1.Overview: Automated ETL pipeline (Gmail → n8n → Supabase PostgreSQL) that loads sales data arriving as emailed CSV attachments straight into a database, removing the manual download-and-import step.
2.Workflow: A Gmail Trigger watches the inbox for emails with “sales” in the subject and downloads the attachment; Extract from File parses the CSV into rows using the header; Insert Rows writes them to the sales table.
3.Tools: n8n Cloud for orchestration, Gmail, PostgreSQL on Supabase with the schema created via SQL, Git and GitHub.
4.Data: Sales order records with OrderID, order date, customer name and email, amount, region and status, with explicit column mapping for the OrderDate field.
5.Result: 10 new orders were appended automatically from a single email, growing the table from 20 to 30 records with zero manual steps.
6.Impact: Low-code, hands-off and repeatable ingestion that keeps the sales database up to date, automatically.

Project Details

**Project info (about 1,000 characters):**

> Unstructured-to-Structured Invoice Data Pipeline is an n8n automation that removes manual invoice data entry. Invoices in PDF or image format are placed in a Google Drive folder, and one click starts the workflow. It loops through each file and routes it by type with a Switch node. Digital PDFs are read with the Extract from File node, while images and scanned files go through the OCR.space API. The raw text is then passed to an Information Extractor node powered by Google Gemini, which uses a JSON sample to return consistent fields such as invoice number, date, billing company details, customer name, tax and total amount. Results are saved in two linked Google Sheets: Invoice Details for header data and Invoice Items for line items, one row per item. A UID-based match prevents duplicate records when files are reprocessed. The result is a fast, accurate, repeatable pipeline that saves hours of work.

Project Details

AI-Powered SQL Analyst | Natural Language to PostgreSQL

Overview:
An n8n AI Agent that lets users ask business questions in plain English and get answers straight from a PostgreSQL database. No SQL knowledge needed.

How it works:

A chat message triggers the AI Agent.
Google Gemini understands the question and plans the answer.
The agent checks the database schema, writes the SQL query and runs it using Postgres tools.
Results are explained back in chat, while Simple Memory keeps the conversation context.

Key features:

1.Natural language to SQL conversion
2.Schema-aware query generation
3.Follow-up questions for unclear requests
4.Clear fallback message when data isn’t available
5.UAT-tested on complex business questions

Tech stack:
n8n, Google Gemini, PostgreSQL, Simple Memory

Outcome:
Teams get instant insights without depending on a technical person, which saves time and removes reporting bottlenecks.

Project Details

AtliQ AI Agent is an autonomous sales-monitoring agent built during the Codebasics Data Analytics Bootcamp. At AtliQ Hardware, stakeholders checked dashboards by hand to spot sales drops, then waited for an analyst to explain them, so issues surfaced late. This agent automates the whole loop: detect, investigate, explain, notify. Every morning at 8 AM, a scheduled Python job checks only complete days, comparing net sales with the median of the same weekday over four weeks, and flags customers who suddenly stopped ordering. When a drop is found, an investigator runs safe, read-only SQL to drill down by customer, product, region and channel, and an LLM (Groq) writes the likely cause. The right owner gets a Gmail alert with a Word summary attached, escalating one level up if the issue persists. Each week, a one-page digest goes to all stakeholders. Every run is logged to a run_log table for auditing. Stack: Python, Groq, SQL, Gmail, Task Scheduler.