I work with real business data (sales, cost, footfall, travel spend) and follow it through to a recommendation someone has to accept or reject. Open a project to read the overview, the full report, or the deliverables.
I'm an Operations Analyst based in Dubai, working where the data is rarely tidy and the decisions are real. I'm building toward a career in strategy and management consulting, and I approach projects the way a good consulting engagement is scoped: start with a business question that can't yet be answered, and finish with a recommendation that drives action.
Alongside my work, I'm a professional football player, which has shaped how I think about performance, discipline, and decision-making under pressure. Whether on the pitch or at work, I like solving hard problems with data and structured thinking.
I'm looking for analytics, operations and strategy projects where I can solve real business problems, build practical solutions and make a measurable difference. That could mean working with businesses, joining a consulting engagement, or taking on a difficult data problem. I'm most interested in work where the insight changes a decision.
My football started at the Fursan Hispania academy in 2014. I spent four years in England, at the Fleetwood Town FC Academy and then Rossall School, before playing for Shabab Al Ahli in Dubai. In August 2026 I moved to Al Ittifaq FC and signed my professional contract.
First-team player in the UAE First Division, after moving from Shabab Al Ahli in August 2026.
Describe what happened, decide what to do, predict what comes next.
BrewHaven's owner felt business had gone flat but couldn't say why. The engagement ran in three connected parts, each asking a harder question than the last as the data got messier.
Why a fictional company. BrewHaven is invented so I could run the full analyst workflow on data I understood completely: realistic seasonality, cost structures and footfall behaviour, with no real business involved. I treated it like a client engagement so the focus stayed on turning numbers into a decision. Every figure below comes from the underlying data, and each method is explained so the reasoning can be checked.
Part I: Descriptive
The owner, Daniel, felt business had gone flat but couldn't say why. He had a year of point-of-sale data across four cafés and had never looked at it properly. The brief: give him a clear read on where the money comes from, and one useful recommendation.
Two files arrived: a transactions table with one row per item, and a small table describing the four cafés. Transactions carried only a store code, not a café name, so step one was a join on the shared key. With that in place, every question below becomes a matter of grouping sales and adding them up.
Revenue is heavily uneven. Downtown, the flagship, took ~£37,600 over the year. Riverside, the newest café, took ~£13,300, about a third of the flagship and well behind the rest. Worth flagging, but revenue alone doesn't tell you whether a café is healthy.
Splitting by month sharpens it. Downtown and Uptown are steady to rising, Campus swings with the university term, and Riverside drifts down almost every month. A café that is both the smallest and the only one falling is the obvious place for attention, though the data can't yet say why.
Three moves and one caution. Weight staffing toward the morning peak. Keep coffee central to menu and promotions. Don't cut tea yet: low revenue isn't low profit. And don't close Riverside on low revenue alone: there was no cost data yet to say whether it truly loses money. That caution set up the next engagement.
Part II: Diagnostic & decision
Daniel took the advice not to close on a hunch and came back with the financial and footfall data to settle it. The question: keep Riverside or close it, and if keep, what has to change. The work moved from revenue to profit, a far more honest measure.
The data came from three systems and didn't line up, which is normal and part of the job. Sales dates arrived in a mixed spreadsheet format and had to be parsed carefully. The accounting file held numbers as text with stray pound signs and thousands commas, plus a couple of blank utility bills filled with each café's own median. The door-counter file had duplicates and impossible readings (blanks, a zero, a negative count), treated as counter faults and excluded. One product cost was missing and another misspelled, so a small slice of sales couldn't be costed. That gap was measured and reported, not hidden.
Bringing sales and costs together changes the Part I picture. Downtown, Uptown and Campus all profit. Riverside loses ~£2,800 across the year. So the low-revenue café is indeed the problem, but that still doesn't explain the cause.
That distinction is the difference between a café worth fixing and one worth closing.
Here the Part I instinct about tea gets tested and fails. Tea is smallest by revenue but highest by margin (~86%), ahead of coffee (76%). Sandwiches, healthy on revenue, are thinnest at 39%. Cutting tea would have thrown away the most profitable line on the menu. Rank by profit, not sales, before touching the range.
The door counters answer a different question: is Riverside short of visitors, or short of buyers? Its conversion is lowest of the four at 42%, and spend per visitor is lowest too. People come through the door, but fewer buy, and those who do spend less. These are fixable operational levers, not a dead location.
Keep Riverside open and run a time-limited turnaround, conditional on renegotiating the rent. Modelling a 40% rent cut alongside a lift in conversion to the chain average takes Riverside from a £2,800 loss to a £3,700 profit. Rent is the biggest single lever; conversion is second. The one risk worth naming: the lease runs to 2029, so the saving depends on the landlord agreeing. If they won't move, the recommendation changes.
Part III: Forecast & investment
With the turnaround underway, Daniel started asking about the future: a view of the year ahead for cash and staffing, and help placing a £20,000 budget across five competing ideas.
Three years of monthly data show revenue rising from ~£82,000 (2023) to ~£102,000 (2025). Steady growth is the backdrop for any forecast.
December runs ~26% above average, August well below, with Campus swinging hardest as the student trade empties over summer. This is the calendar to plan cash and rotas around.
Each café's history is deseasonalised, a straight-line trend fitted to the smoothed series and projected twelve months out, then the seasonal shape re-applied so the forecast keeps the same December peak and August dip. Each café is forecast on its own and summed.
The combined 2026 forecast is ~£104,000. One caveat: Riverside has only two years of history and was declining, so its forecast is far less reliable than Downtown's. For a short, unstable series the sensible move is to show a range and state the assumption. A forecast is a judgement to be owned, not a number the computer settles.
Five options, each with an upfront cost, expected yearly benefit and useful life. Payback is quick but ignores everything past break-even. Simple ROI is lifetime profit over cost. Net present value discounts every future year back into today's money. Positive creates value, negative destroys it. NPV drives the decision, at the agreed 10% discount rate.
With £20,000 to spend, the best combination is the three positive options: the Riverside turnaround, a second espresso machine at Downtown, and a delivery partnership. They cost exactly £20,000 and produce a total NPV of ~£21,600. The clearest lesson of the whole project sits here: the biggest, most exciting project can be the worst investment, and a quick payback is not the same as a good one.
The three-part write-up is the analyst's working. The dashboard is what I'd actually hand Daniel. I pulled the same sales, cost and footfall files into Power BI and modelled them as a star schema, with one store table feeding two fact tables, so the totals stay consistent however he slices them. I cleaned the messy parts in Power Query, including a stray pound sign in the rent column and a cost file that only covered 2025, then wrote DAX measures for revenue, gross profit, margin, year-on-year growth, and revenue per visitor.
It has three pages: an executive summary with the headline KPIs and the monthly trend, a store deep-dive with a month-by-month table, and a forecast page projecting the next twelve months with exponential smoothing.
The dashboard surfaced the thing raw revenue was hiding. BrewHaven Riverside is the smallest café by sales at roughly £26K. It's the obvious one to close on a revenue chart. But it converts every visitor at £1.41, the highest of any location and about four times Downtown's £0.34.
That one comparison flipped the recommendation from "shut Riverside" to "work out what Riverside does well and copy it across the other three."
The outputs from the engagement: the written report and the Power BI dashboard, plus the full chart pack below.
The complete three-part write-up with method, figures and recommendations.
Read in browser →Three pages (executive summary, store deep-dive and a 12-month forecast) modelled as a star schema with DAX measures.
See the dashboard →All thirteen figures produced for the engagement, in Python with matplotlib.
Hypothesise first, then test it against 526,000 real transactions.
Dubai residential prices rose roughly 60 to 80% between 2021 and early 2026, and rents in previously "affordable" communities jumped more than 20% in a single year. Government and industry are asking the same question: is the market maturing into something more affordable and sustainable, or is affordability still getting worse? And where should a housing authority act?
Why real government data, presented this way. This is a self-directed analysis of public data, not client or government work, and no department commissioned it. I chose it because it's live, real-money policy territory: the kind of market-and-policy question a public-sector advisory team gets hired to answer, run here on the actual Dubai Land Department transaction record rather than a simulation. Every number below comes from that data, and I explain every call I made so it can be checked.
Approach
Before opening the data I wrote down what I expected to find: that prices would have risen sharply and then flattened in 2026, and that unit sizes would be shrinking as developers built faster and smaller to maximise volume. Stating that up front means I was testing a hypothesis against the data, not fishing for a story afterwards.
The question broke into five sub-questions, each becoming a section of the analysis and a page of the dashboard: is the trend up or down, where is it moving fastest, which unit types are shifting, how does each area compare to the mid-market, and can the trend be forecast.
The Dubai Land Department's open transaction export runs from 1966 to July 2026. It is two files, about 1.1GB and 1.75 million rows combined, with every field duplicated in Arabic and English. The first job was trimming it to the question: sales only, not mortgages or gifts; residential units and villas; from 2023 onward. That left 526,128 rows, with a median price of AED 1.4M and median AED 17,400 per square metre. Community names in the raw data are official Dubai Municipality codes, not the marketing names: Marsa Dubai is Dubai Marina, Burj Khalifa is Downtown Dubai. Those were mapped to names a reader would recognise before anything else happened.
Q1: Do prices rise or fall?
Median price per square metre climbed from about AED 15,240 in 2023 to AED 18,137 in 2025, a rise of roughly 19%, before moving sideways through 2026, roughly AED 17,600 to 19,100 a month. Transaction volumes, which grew strongly across 2024 and 2025, cooled in the first half of 2026 too.
Median unit size fell over the same window, from about 84 square metres in 2023 to 77 in 2025, dipping close to 70 by mid-2026. Buyers are paying more per metre for a smaller home.
Q2: Where prices move fastest
Prices diverged sharply around a citywide median near AED 18,600 per square metre in 2026. Prime areas such as Palm Jumeirah and Nad Al Shiba First traded 100 to 115% above that benchmark, while a high-volume affordable belt (Al Warsan First, Dubai Investment Park, Wadi Al Safa) sat 15 to 52% below it.
Q3: Unit types and sizes
Studio price per square metre rose about 30%, from AED 14,386 in 2023 to AED 18,766 in 2026, even as the median studio shrank below 37 square metres. Off-plan sales made up 69% of all transactions, pointing to an investor-led pipeline weighted toward compact units.
Q4: The mid-market benchmark
Defining the benchmark as the city-wide median AED/sqm each year makes the squeeze visible directly: the gap between prime communities and the mid-market benchmark widened every year of the series, the clearest single signal of where affordability is being lost fastest.
Q5: Forecast
The forecast assumes the recent pattern holds and can't see shocks such as interest-rate moves, new supply coming online or wider regional events. It's a simple, defensible projection, not a prediction to be taken on faith, and the caveat sits next to it rather than being left implicit.
An affordable-housing department should focus incentives and land release on adequately sized end-user homes, rather than sub-37 square metre studios, in the below-benchmark communities that already carry genuine demand (Wadi Al Safa 5, Al Hebiah Fourth and Dubai Investment Park among them), where per-metre prices remain within reach of mid-market buyers. The natural next step is Ejari rent-contract data, to test whether renting is easing or tightening alongside sale prices, and mapping the off-plan supply pipeline by unit size so future policy can steer stock toward liveable, affordable footprints instead of investor-grade micro-units.
The outputs from the engagement: the findings memo and the five-page dashboard.
One page, answer-first: the headline, what the data shows, the caveats, and the recommendation.
Read in browser →Five pages, one per sub-question, built on the aggregated output tables.
See the pages →All five pages of the Power BI dashboard, one per sub-question.
Weather windows × the project pipeline, from public data alone.
When and where is offshore installation activity likely to peak across the Gulf and wider MENA region over the next few years, and what does that mean for marine warranty survey (MWS) demand? How should an energy-and-marine consultancy plan its surveyor resourcing and business development around it?
MWS work is triggered by offshore operations like load-outs, transport and installation. Those operations only go ahead in workable weather, and the projects that need them are set by the construction pipeline. Looking at the seasonal weather windows and the pipeline together shows when survey demand will spike, when it will go quiet, and where the growth markets are.
Why public data only. ABL Group is an energy and marine consultancy whose Gulf work centres on marine warranty survey, the independent review and approval of high-risk offshore operations such as load-out, transport and installation. This report uses only publicly available data and contains no ABL Group internal information, client data or pricing. It's a self-directed analytical exercise, safe to share externally and useful internally. The analysis is deliberately hypothesis-led: expectations were written down before the data was examined, then tested against it.
Part I: Arabian Gulf
The starting assumption was that weather windows would limit Gulf operations much as they do the North Sea. The data did not support that.
Four years of ERA5 reanalysis show the Gulf is a calm, shallow sea with a median significant wave height of just 0.38 m, and even the 90th percentile stays near 1 m, edging up only modestly in winter. The largest wave in four years was 2.58 m.
At a routine 1.5 m limit, the Gulf is workable about 98% of the year, dipping only to around 90% in February, so there is effectively no seasonal constraint. The picture changes as the limit tightens. At 0.75 m, which represents a wave-sensitive heavy lift, a clear season emerges: operability runs near 95 to 98% from July to October but falls to about 60% in January and February.
The region holds 151 offshore oil and gas fields. Most are already producing, but 27 sit in the statuses that generate survey work: fields in development, newly discovered, or being decommissioned. These are concentrated in the UAE and Egypt, followed by Kuwait, Iran, Qatar and Saudi Arabia.
Offshore wind, the fastest-growing source of offshore survey work globally, has barely reached the Gulf. The region has just 3 offshore wind projects against more than 2,000 worldwide, led by China, Vietnam, South Korea and the United Kingdom. For now, offshore wind is not a Gulf market. If ABL wants to follow that demand, the signal points to Asia-Pacific and Europe.
Resource steadily through the year. Routine Gulf survey work is rarely stopped by weather, so plan surveyor capacity around the project pipeline rather than a seasonal window. Time the most wave-sensitive heavy lifts into the late-summer window, roughly July to October, and keep them clear of January and February. Focus business development on the 27 survey-active offshore fields, concentrated in the UAE, Egypt, Qatar and Saudi Arabia. Treat the offshore wind build-out in Asia-Pacific and Europe as the growth play, since the Gulf does not yet have one.
Part II: Resourcing across ABL's regions
An earlier study showed the Arabian Gulf is calm enough to work almost year-round, so survey demand there is set by the project pipeline rather than the weather. This report asks the natural follow-on question for a global business: if the Gulf is unusually steady, how does offshore workability differ across ABL's other regions, and what should that mean for how each office is resourced? ABL operates from hubs including Dubai, Aberdeen, Houston, Singapore, Rio de Janeiro and Perth, each on a different body of water with its own weather.
The approach extends the same operability logic. The Gulf is quantified directly from four years of ERA5 reanalysis. The other five hubs are classified from well-documented seasonal patterns (the Atlantic hurricane season, the Australian and North-West Shelf cyclone season, the Asian monsoons, and North Sea winter storms), each rated good, marginal or poor for a given month. Mixing one measured region with five documented ones is deliberate for a first pass: it gets the strategic picture in place quickly and honestly, and names the next step, which is running the ERA5 pipeline for every regional box so each curve is measured rather than described.
The hubs behave very differently. The Gulf is green across the year. The North Sea opens in late spring and closes in late autumn. The Gulf of Mexico is the near-mirror image: good in winter and spring but hit by the hurricane season in late summer. Brazil and North-West Australia favour the southern half of the year, and the South China Sea is shaped by the monsoons.
Because the seasons are offset, the hubs rarely close at the same time. In almost every month at least three or four of the six have a good window while others are shut. For a single office that's just local weather; for a global business it's an opportunity, because demand that goes quiet in one region is often picking up in another.
The practical idea that falls out of the calendar is to stop staffing each office only for its own peak and instead manage a shared pool that moves toward whichever window is open. Across the northern summer, the Gulf of Mexico is quiet under the hurricane season while the North Sea is at its busiest, so effort can shift from Houston toward Aberdeen. Across the northern winter, the North Sea closes while the Gulf stays open and Brazil enters its best months, so effort can shift the other way. The Gulf sits underneath all of it as a stable base that does not need to flex. None of this removes the need for local presence, certification and relationships. It's a way to smooth the peaks and troughs at the margin.
Anchor on the Gulf: resource Dubai and Abu Dhabi to a steady base-load, since it needs little seasonal adjustment. Staff the seasonal offices to their window: size Aberdeen, Houston, Rio and Perth to their peak months and plan the off-season as deliberate slack for training, leave and mobilisation. Rotate a shared surveyor pool across offsets, starting with the clearest pair, Houston and Aberdeen. And measure before committing: quantify each region with the ERA5 pipeline so resourcing decisions rest on measured operable-day curves, not documented patterns alone.
The two written reports are the analysis. The dashboard is the operational tool: three pages built in Power BI on the same operability and pipeline outputs, for flipping through month by month rather than reading top to bottom.
The seasonal shape from the report becomes something a resourcing manager can filter and read at a glance, rather than a static chart in a document.
The outputs from the engagement: two written reports and the Power BI dashboard, plus the chart pack below.
The primary report: method, five figures and the resourcing recommendation for the Arabian Gulf.
Read in browser →Extends the same method across ABL's six global hubs, including the North Sea, Gulf of Mexico, Brazil, North-West Australia and Singapore.
Read in browser →Three pages: the weather window, the pipeline and the demand outlook.
See the dashboard →All seven figures from both reports.
Revenue outlook, valuation, and whether the money buys success.
Manchester United is a football club and a company listed on the New York Stock Exchange, which makes it one of the few clubs whose accounts are genuinely public. This study treats it as a business through three linked questions: where is revenue heading, what is the club worth, and does the money spent on wages buy the results it's meant to? The club reports three streams (Commercial, Broadcasting and Matchday), and this builds on those filings plus public wage-bill and league-table compilations.
Why this, and why now. This is an independent analysis of public data, not client or club work, and no one commissioned it. Manchester United plc's NYSE listing means its revenue, costs and debt are disclosed in SEC filings (Form 20-F, 6-K) rather than estimated, which makes it one of the few football clubs where "the business behind the badge" can be examined directly rather than inferred. Revenue figures below are compiled from public reporting and investor releases and should be checked line by line against the filings. The wages-to-points result uses a small starter sample, so it illustrates the method rather than settling the question. The valuation rests on stated assumptions for margin, discount rate and multiple, so it is a defensible range, not a single true number. None of this is investment advice.
Part I: Revenue
In 2024/25, revenue reached a record £666.5m, with Commercial income (sponsorship and retail) making up roughly half of it. The more interesting number sits in Broadcasting: it fell by roughly £49m that year for one reason only: the team played in the Europa League rather than the Champions League. Commercial and Matchday growth is measured from 2022 onward, to avoid the COVID-distorted years, and capped at a sensible ceiling because matchday income is bound by stadium capacity rather than free to compound. Broadcasting is projected under two scenarios, with and without Champions League football, because that qualification is the real driver, not a trend line.
The forecast splits accordingly: about £746m in 2026 with Champions League football against about £696m without it, and the two paths widen further out to 2028. For a club that posted a loss in a record-revenue season, that swing is the difference between profit and loss.
Part II: Valuation
Two lenses give very different answers. A discounted cash flow, built from the forecast revenue and a free-cash-flow margin, supports only about £0.75bn of enterprise value, which after roughly £730m of net debt leaves almost nothing for shareholders. The multiples view (the four-to-six-times-revenue range at which clubs actually change hands, benchmarked against published valuations) tells a completely different story, putting the club at roughly £2.7bn to £4bn, in line with Forbes and Sportico's published figures.
Both figures are defensible; the gap between them is the point. Investors are paying for a global brand and a scarce asset, not for the cash the club actually generates.
Part III: Wages vs. points
Across the league, wage spend usually explains most of where a club finishes. League points were regressed on wage bills across clubs and seasons, and each club's residual shows whether it beat or fell short of what its pay predicts. On the sample here, United sits clearly below the fitted line: it has finished lower than its wage bill would predict. This is the quantified version of a familiar complaint, and it's the on-pitch side of the same coin as the financial findings above.
Bringing it together
The three workstreams converge on the same thing. The biggest uncertainty in the revenue forecast, the swing between a modest and a strong valuation case, and the shortfall between spending and results all trace back to Champions League qualification. It is the largest financial lever the club has, and exactly what a wage bill among the highest in the league is supposed to secure. For United, getting back into Europe's top competition is as much a commercial strategy as a football goal.
The revenue figures should be checked line by line against the annual filings rather than taken as final. The wages-to-points result needs the full set of clubs and seasons before it's conclusive rather than illustrative. The valuation should be stress-tested across a wider set of margin, discount-rate and multiple assumptions so the range is fully explicit. Each step would move this from a first pass toward something publishable. This is independent analysis of public data, not investment advice.
The outputs from this engagement: the full report and its four figures.
The full report: revenue forecast, two-lens valuation, and the wages-to-points result, with caveats stated up front.
Read in browser →All four figures from the report.