Skip to content
DATAVERSE

The work

Proof beats promises.

This is the kind of work you finish here. Four capstones and twelve case studies, all yours to show.

01

A dashboard a manager would use

Delivery times for a food app, cut by hour and area, with the one chart that explains the problem.

See the program
Late orders · weekly view
Orders12,48012,480
On time81%81%
Worst hour8 to 9 pm8 pm
1012268 pm10
02

A model you can explain

A notebook that predicts who is about to cancel, with the score and what it means.

See the program
churn_model.ipynb
# who is about to cancel?
from sklearn.ensemble import RandomForestClassifier
model = RandomForestClassifier(n_estimators=200)
model.fit(X_train, y_train)
print(model.score(X_test, y_test))
Accuracy on unseen viewers0.87
03

An agent that finishes a task

A support agent that reads an order, checks the policy and decides, step by step.

See the program
Refund agent · run 14
  1. Goal“My order came 50 minutes late. Can I get a refund?”
  2. Look up the orderget_order(id) delivered 52 min after the promised time
  3. Check the policysearch_policy("late delivery") over 45 min qualifies
  4. DecideRefund approved. Amount is above the limit, so a person confirms it.
04

A recommendation with numbers

A one-page memo that picks an option and shows why.

See the program
Decision memo · next dark store

Open the next store in the north zone.

North zone86
East zone64
Old city41

Highest unmet demand, shortest ride from the warehouse, rent within budget. Score out of 100 across the three.

Illustrations of capstone work, drawn with practice data. Not screenshots of client or company systems.

Twelve problems worth solving.

Three per program. Each one ends in something you can show.

  • SwiggyData Analytics

    Why do some orders arrive 15 minutes late?

    Dig through delivery data to find where the time goes, and show it on one dashboard.

    • SQL
    • Power BI
    • Storytelling
  • MyntraData Analytics

    What actually sells during a sale?

    Compare sale weeks with normal weeks and find which discounts earned their keep.

    • Excel
    • SQL
    • Pandas
  • OlaData Analytics

    Where should drivers wait at 9 am?

    Map ride demand by hour and area, then turn it into a recommendation a city manager could use.

    • Python
    • Tableau
    • Gen AI
  • NetflixData Science

    Who is about to cancel?

    Build a model that spots viewers drifting away, and explain what it found.

    • Python
    • scikit-learn
    • Evaluation
  • SpotifyData Science

    What makes a song a hit?

    Explore track features and test how well they predict popularity.

    • Pandas
    • Regression
    • Visualisation
  • PhonePeData Science

    Is this payment a fraud?

    Train a classifier on lopsided transaction data and decide what a false alarm really costs.

    • Classification
    • Metrics
    • Python
  • ZomatoAgentic AI

    Can an agent handle a refund request?

    Build a support agent that reads the order, checks the policy and decides, with a human as backup.

    • Tool calling
    • RAG
    • Guardrails
  • MakeMyTripAgentic AI

    Can it plan a weekend in Goa on a budget?

    A planning agent that searches, compares and puts together a trip that fits the money you have.

    • Multi-agent
    • Planning
    • APIs
  • FlipkartAgentic AI

    Can it answer ten thousand product questions overnight?

    A catalogue assistant that answers from product documents and says so when it doesn't know.

    • RAG
    • Vector search
    • Evaluation
  • ZeptoBusiness Analytics

    Which dark store should open next?

    Weigh demand, distance and cost to make a recommendation a manager could sign off.

    • Excel
    • KPIs
    • Decision memo
  • NykaaBusiness Analytics

    Why did repeat orders drop?

    Split customers into groups by when they joined and find where loyalty leaks.

    • SQL
    • Cohorts
    • Power BI
  • boAtBusiness Analytics

    Should the next launch be cheaper or better?

    Read pricing and review data, then back one option with numbers.

    • Excel
    • Pricing
    • Storytelling

Swipe for all 12

Learning case studies built on public information and practice datasets. DATAVERSE is not affiliated with these companies.

Ready when you are.

One conversation, no pressure. Tell us where you are and we'll tell you honestly where to start.