Data Analyst and Data Scientist · 8+ Years

Where data meets
business decisions

I translate complex datasets into strategic decisions that move the needle, from churn models with 86% accuracy to supply chain automation that saves 50% in processing time. I work across the public sector, financial services, e-commerce, and startups at the intersection of analysis, machine learning, and commercial impact.

📊
8+
Years of experience
💰
£3M+
Business value delivered
🎯
5
Industries covered

Analyst. Data Scientist.
Stakeholder partner.

I started my career building predictive models and extracting data insights for a digital networks company in Lagos, and quickly learned that the hardest part of analytics is not the model. It is earning the trust of a director with competing priorities and getting them to act on what the data is saying.

Over eight years, I have led analytics projects across public sector waste management, real estate, financial services, and early-stage startups. My work has directly influenced operational efficiency, cost reduction, funding strategy, and machine learning deployments, always with a focus on the decision at the end of the data, not the data itself.

I hold an MSc in Data Science from the University of Salford and am currently based in Manchester. I am open to senior analyst, lead analyst, and data science roles where I can continue building high-impact analytical capability.

Query and data engineering
SQL (MS SQL, MySQL, PostgreSQL) Python R PySpark Hadoop
Machine learning and modelling
scikit-learn Pandas NumPy SciPy Regression Clustering Time Series
Visualisation and BI
Power BI Tableau Google Looker Excel / VBA Matplotlib ggplot
Cloud and automation
AWS Azure Google Cloud Power Automate Microsoft Forms

5 projects · Real outcomes

01
Revenue

Customer churn prediction model — credit card segmentation

Led end-to-end churn analysis for a 10K+ customer credit card issuing project, identifying that 45% of selected customers are awarded the wrong credit cards, which is at-risk of accounts shared, and three early behavioural signals invisible to the sales team. Partnered with customer support and operations teams to redesign the intervention, shifting from reactive to predictive outreach 45 days before renewal.

Business problem
-The business had no early warning system for customer churn and no reliable way to interpret customer sentiment from large volumes of unstructured feedback data.
My role
- Lead Data Scientist. Owned the full model lifecycle from data extraction and cleaning through to deployment, documentation, and stakeholder communication.
Python (scikit-learn) SQL Tableau Salesforce Mixpanel
Churn reduced 32% within two quarters · £425K ARR retained · CS coverage improved 3× with no extra headcount
Visualisation specification
Charts built
(1) KNN classifier (Minkowski) showing churn probability by customer category to implement the prefect card type for each customer.
(2) Histogram chart of the top 15 feature importances from the customer demographics.
(3) combined histogram chart: comparing the income of various customers based on their respective ages.
All models were tested and trained, and the overall accuracy was 93%.
02
Operations

Applying association rule mining - market basket analysis

Designed this transactional analysis, which contains all the purchases occurring between, integrating lead-time variance, customer shopping and inventory, buffering into a single operation. where the model monitors the purchase of goods in wholesale and services rendered online, quantifying patronage frequency.

Business problem
- Inventory analysis caused the detection of 18 stockout events in one fiscal year, costing £187,000 in lost sales and damaging key retail partnerships.
- Identifying the frequency of consumers and the most popular item in a purchasing basket.
My role
- Lead analyst and data model owner. Collaborated with procurement, logistics, and finance to define risk dimensions and thresholds.
Python (pandas) SQL (BigQuery) Power BI SAP ERP dbt
Stockouts cut 25% · £46,750 saved annually in analyst hours · Ddashboard displays the clients who make the most repeat purchases.
Visualisation specification
Charts to build
(1) Bar chart showing the most popular item at a specific moment .
(2) 12-month trend line: stockout frequency month-over-month with model deployment marker.
(3) Drill-through table: top 10 at-risk suppliers with current score, trajectory, and recommended buffer days.
03
Pricing

Statistical analysis — factors affecting economic labour

Inflation, import and export of goods, gross capital creation, industry, manufacturing industries, and other elements that impact economic labour in the listed countries are included in the statistics.

Business problem
- Pricing was volume-based. No framework existed to evaluate margin trade-offs at the SKU level, leading to chronic under-pricing in high-demand categories.
My role
- Lead analyst. Defined the methodology, delivered the final board-ready recommendation deck.
R (statsmodels) Excel (scenario model) Tableau
Gross margin up 4.2pp in pilot categories · £800K+ incremental margin in Year 1 · Adopted as the standard pricing process company-wide
Visualisation specification
Charts to build
(1) Four-quadrant scatter: price elasticity coefficient (x) vs. current gross margin (y), with bubble size = revenue volume — reveals "safe to raise" vs "hold" quadrants.
(2) Waterfall chart: margin impact of proposed price changes by category.
(3) Before/after conversion rate line chart for pilot SKUs across a 90-day window.
04
Risk & Compliance

Fraud pattern detection — financial services transaction monitoring

Built an anomaly detection pipeline processing 2M+ daily transactions to surface high-risk behavioural clusters that existing rule-based filters consistently missed. Worked directly with the Chief Risk Officer and external auditors to ensure the model met regulatory documentation standards before production deployment.

Business problem
- Rule-based fraud filters had a 2.1% false-negative rate. Regulatory pressure required a demonstrably improved detection methodology within two quarters.
My role
- Lead analyst and model author. Sole data science contributor — coordinated with risk, compliance, and engineering teams throughout the project lifecycle.
Python (Isolation Forest) Spark SQL Databricks Power BI Confluence
False negatives down 38% · £3.1M in fraud losses avoided in first 6 months · Regulatory audit passed on first submission
Visualisation specification
Charts to build
(1) Side-by-side confusion matrix comparison: rules-only baseline vs. model-assisted detection.
(2) Time-series of daily flagged transactions with confirmed fraud overlay — shows signal-to-noise improvement post-deployment.
(3) Anomaly score distribution histogram with decision threshold annotated.
05
Marketing

Text Mining — international hotel reviews

I led an end‑to‑end analysis of a large‑scale hospitality review dataset (53,647 customer reviews across hotels and restaurants in Thailand and surrounding regions). The dataset included five structured fields—ID, review date, location, establishment name, and full review text—each containing unique customer insights.

Business problem
- Lack of insight into customer sentiment and Service Perception
- Inability to benchmark performance across locations and Star Ratings
My role
- My work focused on extracting key phrases, identifying sentiment patterns, and uncovering service‑quality themes that matter most to guests. This enabled hotels, particularly 2‑ to 3.5‑star establishments concentrated along the Phuket coastline, to understand what customers valued and where improvements were needed.
Python (PyMC) SQL Google Colab Looker Google Analytics 4
ROAS improved 22% · £1.8M budget reallocation approved · Model embedded into annual operations / customer service planning cycle
Visualisation specification
Charts to build
(1) Word Clouds: Text data was visualised to highlight the most frequently occurring terms in customer reviews.”
(2) generating actionable insights for service improvement
(3) operations budget flow before vs. after reallocation recommendation.

Technical leadership &
process improvement

Technical leadership
  • Designed and deployed a Power Automate workflow at Lancashire County Council that automated daily report distribution to 20 or more stakeholders, replacing an entirely manual process
  • Architected the vendor scoring algorithm at Marketbuddyng, which was adopted directly into the product roadmap for vendor financing partnerships
  • Integrated data from 5 or more APIs and databases at M and E Realty into a unified reporting layer, reducing fragmentation and improving decision quality across finance
  • Built predictive models achieving 91 to 93% accuracy in both churn forecasting and revenue prediction, deployed to production and used by business leadership
  • Enhanced legacy systems for accessibility by staff with limited IT experience, reducing friction and support burden across the team
Training and process improvement
  • Delivered staff training programmes on new and existing systems at Lancashire County Council, resulting in a 70% KPI improvement across the team
  • Identified and implemented continuous improvement opportunities across operational teams, increasing work rate and turnaround time by 20%
  • Streamlined multiple complex data sources into efficient formats integrated with Microsoft Forms, reducing manual data entry errors
  • Distilled complex technical findings into accessible instructions and presentations for both non-technical staff and senior stakeholders
  • Supported the development and implementation of new council policies through data evidence and operational analysis

Recruiter one-pager

Mobolaji Joachim Anthony
Data Analyst and Data Scientist
@ mobolajianthony2805@gmail.com
T (+44) 07721 665 223
L Manchester, United Kingdom

Data Scientist and Analyst with 8 years of experience spanning public sector operations, real estate, financial services, and early-stage startups. Specialises in turning complex datasets into decisions that drive revenue, reduce cost, and improve operational performance. Proficient across the full analytics stack from SQL and Python modelling through to Power BI and Tableau dashboards presented directly to executive stakeholders. Holds an MSc in Data Science from the University of Salford.

01
Waste Operations Analytics and Automation, Lancashire County Council
Analysed 29,000 waste haulage operations worth 3,000,000 GBP. Built a Power Automate workflow that replaced manual daily report distribution for 20 or more stakeholders and standardised data accuracy to 97 percent.
Manual processing time cut 50% and operational throughput improved 15%
Tools: Power BI, Excel, Power Automate, SQL, Microsoft Forms
02
Compensation and Cost Analytics, M and E Realty
Built SQL-based reports analysing compensation totalling 230,000 GBP for over 100 employees. Audited shipping spend and integrated 5 or more data sources into a unified reporting layer. Delivered predictive models to 87 percent accuracy.
79,000 GBP in annual savings identified and reporting time reduced 30%
Tools: SQL, Python, Tableau, Power BI, QlikView
03
Customer Churn Prediction and Sentiment Analysis, Krystal Digital Networks
Built a machine learning churn prediction model at 86 percent accuracy and a text pattern algorithm that improved customer sentiment analysis accuracy by 12 percent. Insights shaped the organisation approach to school-sector client engagement.
Churn model deployed to production and adopted into business strategy
Tools: Python, scikit-learn, Pandas, NumPy, Power BI, Tableau
Query and data
Python, R, SQL (MS SQL, MySQL, PostgreSQL), PySpark, Hadoop
Machine learning
scikit-learn, Pandas, NumPy, Regression, Clustering, Time Series, EDA
Visualisation and BI
Power BI, Tableau, Google Looker, Matplotlib, ggplot, Excel, VBA
Cloud and automation
AWS, Azure, Google Cloud, Power Automate, Microsoft Forms
MSc Data Science
University of Salford Manchester, 2022 to 2023
BSc Computer Science
ISFOP Cotonou, 2013 to 2017
AWS Academy
Data Analytics Certification
Chartered Management Institute
Management and Leadership
Award
Outstanding Performance, Krystal Digital Solutions

Delivered staff training on new and existing systems at Lancashire County Council resulting in a 70 percent KPI improvement across the team. Identified continuous improvement opportunities that increased operational work rate and turnaround time by 20 percent. Architected a vendor scoring algorithm at Marketbuddyng that was adopted into the product roadmap, and supported a 500,000 USD seed fundraise with financial modelling presented directly to the executive team.

Let's talk about
your data challenges

I am open to senior analyst, lead analyst, and data science roles, full-time or contract. I work best with teams that care about the quality of decisions, not just the quantity of dashboards.