Exploratory analysis
Segmentation, trends, and clear charts that explain why a metric moved.
Hi, I'm
Rushikesh Wagh
Product Analyst · Analytics & Data Science
EDA ML Metrics & KPIs Experimentation
4+ years turning raw data into models, dashboards, and decisions that move product and revenue. From funnel diagnostics to churn and fraud models — I care about the so what for the business.
Shipped Impact
Real numbers from production work — not side projects.
24%
Revenue increase
Dunzo — merchant pricing model
2% → 7%
Credit funnel conversion
FloBiz — funnel optimisation
20%
Churn reduction
ALLEN Digital — at-risk ML model
£17L / yr
Cost saved
Dunzo — 134 → 39 dashboards
245%
MoM lead growth
FloBiz — new lead gen channels
2×
Student renewals
ALLEN Digital — Live Classroom
About Me
I'm a Product Analyst based in Bengaluru, Karnataka with 4+ years of full-time experience in the analytics domain. I have a B.Tech from IIT (ISM) Dhanbad and have worked across startups and growth-stage companies including Dunzo, FloBiz, and ALLEN Digital.
I help teams make smarter, faster, and data-backed decisions. Whether it's building an XGBoost model to identify at-risk students, optimizing a credit funnel from 2% to 7%, or consolidating 134 dashboards down to 39 — I focus on the "so what?" that moves the business forward.
My work spans product analytics, ML-driven predictions, data pipeline engineering (SparkSQL, Airflow, Hevo), and strategic analysis at the founder's office level.
0+
Years Experience
0
Companies
0+
Datasets Analyzed
0+
Awards Won
How I Work
An end-to-end analytics loop — the same rhythm behind my portfolio repos and my product work.
-
01
Frame the decision
Clarify the business question, success metric, and trade-offs (e.g. precision vs recall for fraud).
-
02
Data & quality
Source tables, validate joins, handle missing values, and sanity-check distributions before modeling.
-
03
EDA & hypotheses
Visualize segments, correlations, and leakage risks; turn patterns into testable hypotheses.
-
04
Model or metric logic
Baselines first, then sklearn / XGBoost-style models, with proper splits and imbalance handling when needed.
-
05
Evaluate fairly
Metrics that match the use case — not just accuracy — plus confusion matrices, ROC, and feature importance.
-
06
Ship the story
READMEs, notebooks, or dashboards so stakeholders see the insight and the recommended next action.
Predictive modeling
Classification & regression with scikit-learn — from churn and fraud to cost prediction.
Imbalanced & risky outcomes
SMOTE, recall-focused evaluation, and business-aligned thresholds for rare events.
Dashboards & KPI design
Streamlit and BI-style views that make metrics self-serve and actionable.
Experience
From founder's office strategy roles to product analytics — always at the intersection of data and business impact.
Jul 2024 – Present
ALLEN Digital — Product Analytics
Bengaluru, Karnataka
- Owned the analytics charter for Live Classroom — 20% YoY increase in attendance and 2x improvement in student renewals
- Built end-to-end XGBoost ML model to identify at-risk students, reducing churn rate by 20%
- Architected core data pipeline using Hevo, SparkSQL, and Airflow for ETL workflows
Dec 2023 – Jul 2024
FloBiz — Founder's Office, Strategy & Analytics
Bengaluru, Karnataka
- Improved credit funnel efficiency from 2% to 7% and R/C metric from 80% to 115%
- Drove 245% MoM increase in leads through new lead gen sources and marketing campaigns
- Conducted market research for launch of insurance, personal loans, gold loans, and credit cards
- Produced investor MIS reports, Townhalls, and MBRs
Jun 2022 – Nov 2023
Dunzo — Founder's Office, Analyst 1 → Analyst 2
Bengaluru, Karnataka · 1 year 6 months
- Built merchant pricing model → 24% revenue increase for DMS business
- Led CPT optimization project → 10%+ CPT reduction across the business
- Consolidated dashboards from 134 to 39 → cost saving of ₹17L/year
- Contributed to surge pricing, delayed delivery rollouts with pre/post feature reads
🏆 5 Monthly awards · 2 Excellence awards · 1 Quarterly award
May – Jun 2021
Tata Steel — Business Analyst Intern
Khopoli, Maharashtra
Jul 2019 – Jun 2020
Student Alumni & International Relations Cell — Data Analyst Intern
Dhanbad, Jharkhand
2018 – 2022
IIT (ISM) Dhanbad — B.Tech, Engineering
Dhanbad, Jharkhand
Analytics Co-Pilot
Quick, private utilities that run entirely in your browser — no sign-up, no AI, no server. Useful for experiment readouts, retention tables, and readable SQL.
100% client-side · $0 infraSignificance test
Two-proportion z-test. Enter visitors and conversions for each variant. Large-sample normal approximation — for small samples, treat as directional only.
Sample size calculator
How many users per variant do you need before running the test? Enter your baseline conversion rate, minimum detectable effect (MDE), desired power, and significance level.
JSON formatter
Paste minified or messy JSON on the left. Click Format to pretty-print with syntax highlighting on the right. If the JSON is an array of objects you can also download it as a CSV file.
SQL formatter + keyword hints
Indents SELECT columns one-per-line, AND / OR / ON under their clause, and colour-highlights keywords, functions, strings, and numbers. Heuristic — works well for standard analytical SQL. Copy button outputs plain text.
Keywords detected (reference)
Cohort retention heatmap
CSV format (wide): first column = cohort name; remaining columns = period 0, 1, 2… (counts or retention %). Example header: cohort,0,1,2,3. Values between 0 and 1 are treated as fractions.
Open Source Projects
Public repos with notebooks, EDA, and models — each README summarizes methods, metrics, and business takeaways.
Sales & Revenue Dashboard
Interactive Streamlit dashboard on the Superstore dataset with KPI cards, trend charts, discount-vs-profit analysis, and real-time filters.
Key insight: Sub-categories with >20% avg discount have negative profit margins
Telco Customer Churn Prediction
ML pipeline predicting customer churn using Logistic Regression & Random Forest. Surfaces top reasons customers leave and concrete retention actions.
Key insight: Month-to-month contracts churn 3x more — push annual plans with incentives
E-Commerce Sales Analysis
Customer & Revenue Intelligence Report — deep-dive into an e-commerce dataset with segmentation, revenue driver identification, and growth recommendations.
Key insight: Revenue concentration analysis reveals top customer segments driving 80% of sales
Medical Insurance Cost Prediction
Regression models predicting medical insurance costs — EDA on what drives healthcare spending plus business recommendations for insurers.
Key insight: Smokers pay 3.8x more; obese smokers pay 5x+ vs healthy non-smokers
Credit Card Fraud Detection
ML pipeline tackling extreme class imbalance (0.17% fraud) using SMOTE, Logistic Regression, Random Forest & Gradient Boosting with focus on recall optimization.
Key insight: Accuracy alone is misleading at 99.83% — recall-focused models catch real fraud
Student Performance Prediction
Classification models predicting student pass/fail outcomes using demographic, social, and academic features from the UCI Student Performance dataset.
Key insight: Prior grades, past failures, and alcohol consumption are the strongest predictors of academic outcomes
Skills
Languages
Python
SQL
Data & ML
pandas / NumPy
scikit-learn
XGBoost / Gradient boosting
Imbalanced-learn (SMOTE)
Feature importance & explainability
Jupyter / notebooks
Analytics craft
EDA & hypothesis testing
Metrics design (north-star, funnels)
A/B tests & experiment readouts
Cohort & retention analysis
Dashboards & BI
Streamlit
Plotly
Power BI
Data Engineering
SparkSQL
Airflow
Hevo
Git / Docker
Blogs
Educational content on analytics — from experimentation to ML, metric design, and data storytelling.
Coming Soon
In-depth write-ups on analytics concepts, SQL patterns, A/B testing, cohort analysis, and real-world data lessons are on their way.
Fun
Analytics games for data people — SQL quiz, daily word game & metric blitz.
The Query Gauntlet
Clause & Effect · Syntax Circuit · The JOIN Tribunal
10 questions · 10 minutes · no lifelines. Read real SQL, spot errors, predict output. Enter your name and go — certificate at the finish line.
Question 1 / 10
10:00
🏆 Pro analyst
Congratulations!
You scored . You are operating at professional analyst level on this SQL set.
Your results
Topics to improve
Based on the questions you missed, spend time on:
Copied — paste into LinkedIn if the share window did not open.
Daily · resets at midnight
SQL Word
Guess the 5-letter SQL / data keyword. 6 tries. New word every day.
Blitz · 30 seconds
Metric Blitz
5 KPI & SQL questions. 30 seconds on the clock. No looking it up.
Q1 / 5
0:30