01

Academic Research

★ Master's Thesis  ·  ESIEE Paris, France

A project manager wakes up on day 60 of a 90-day engagement and has no idea whether the project is on track. The warning signs were there in the data. Nobody was looking at them systematically.

Predictive Analytics for Improving IT Project Delivery Outcomes

I built and validated a machine learning pipeline to predict schedule delay in IT software projects using only data available at the project's start — enabling teams to act before delay becomes inevitable, not after it's already happened.

Python Random Forest Logistic Regression Scikit-learn LOOCV
83%Classification accuracy
8.21RF regression MAE (days)
3ML models compared
18Real IT project records
Key finding: Random Forest outperformed Multiple Linear Regression (MAE 8.21 vs 11.90 days) — demonstrating that non-linear models better capture early warning signals in complex project network data.
Academic Endorsement

I have the great pleasure to recommend Madam Rukayat Rauf to any employer. I had the opportunity to be her Professor and her thesis tutor. She is a smart and hardworking person with unparalleled self-drive to excel in her undertakings. She wrote and presented an excellent thesis. Consequently, she had one of the best grades of her class of 2026. For all her qualities mentioned above, I strongly believe that she will have a great career.

Doudou Sidibé Associate Professor, ESIEE Paris, University of Gustave Eiffel | Professor & Master's Thesis Tutor
02

Applied Analytics & BI

A marketing team running campaigns on three channels couldn't tell which one was driving revenue — or why the busiest month wasn't the most profitable.

Power BI · Marketing Analytics

Marketing Campaign Performance Analysis

An interactive dashboard breaking ROAS, conversion, and engagement down by channel, season, and campaign type — so the team could stop spreading budget evenly and start putting it where it worked.

22.5xBest-platform ROAS
63%Revenue from discounts

A finance team spending hours each month compiling a P&L manually — and still delivering it late, with errors that only surfaced in the next cycle.

Power BI · Financial Intelligence

Financial & Income Statement Dashboard

A 4-page executive BI dashboard replacing manual monthly reporting with live visibility — monthly and quarterly modes, one source of truth, zero manual compilation.

4Dashboard pages
ZeroManual data entry

A business treated all its customers the same, wasting marketing budget on churned accounts and ignoring its most valuable loyalists.

Python · CRM Analytics

Customer Segmentation & RFM Analysis

Segmented the customer base using Recency, Frequency, and Monetary value models to identify VIPs, at-risk accounts, and lost customers, allowing targeted retention strategies.

RFMModelling technique
3+Distinct segments found

Sales directors had no visibility into where deals were dropping off, making revenue forecasting impossible.

SQL & Tableau · Sales Analytics

B2B Sales Pipeline Analysis

Analysed CRM data to uncover conversion bottlenecks across the sales funnel, providing sales leadership with a clear view of win rates by rep, region, and product line.

100%Pipeline visibility

Investors need to model potential outcomes based on historical data rather than static snapshots.

Excel · Financial Modelling

NYSE Financial Analysis

Built a dynamic, three-scenario financial model (Bull, Base, Bear) for S&P 500 companies using historical stock exchange data to forecast revenue and operating income.

3Forecast scenarios