Eighteen years of Dubai property data, compared against UAE income growth, in under ten minutes. Codex flags every data issue, waits for your approval before fixing it, and produces an insights document plus a presentation when it is done.
A clean dataset, a written insights document, and a presentation that answers whether the market you are buying into looks anything like 2008. The math is Codex's. The judgement is still yours.
A cleaned price index from January 2008 paired with annual UAE GDP per capita from World Bank data. An insights document covering the periods of price acceleration relative to income growth. A succinct presentation built off the same analysis.
On the Dubai run, prices were 36.9% above the 2008 peak by May 2026. Per capita GDP was only 0.4% higher over the same window. The strongest divergence years were 2013, 2014, 2023, and 2024.
A paid ChatGPT plan with Codex access. Codex runs inside your ChatGPT account. The free tier does not include it. Check what your current plan allows before you start.
A historical price index for your market. Property Monitor for Dubai. Land Registry or Halifax in the UK. Zillow ZHVI or Case-Shiller in the US. CoreLogic in Australia. Monthly, going back to a previous cycle peak.
A per capita income series for the same country. World Bank Open Data is free and downloads as a single Excel file. The UAE series goes back to 1975. Most countries are available on the same portal.
Codex is a coding agent doing market analysis. The math is solid. The conclusion sits on the index methodology you fed it. Read the methodology notes on whichever price series you upload. A rebase or basket change mid-series will skew any cycle comparison unless Codex aligns the segments first.
Two attachments before the prompt. Both are spreadsheets. The price series carries the cycle. The income series tells you whether the cycle is moving with the underlying economy or away from it.
Property Monitor price per sqft, monthly, January 2008 to current. The 2008 start matters. It captures the previous cycle peak, the 2009 crash, the 2013 to 2015 rally, the 2020 reset, and the post-2020 recovery in one continuous series.
Outside Dubai, swap in the equivalent national index. Land Registry or Halifax in the UK. Zillow ZHVI or Case-Shiller in the US. CoreLogic in Australia. Any monthly series stretching back to a previous cycle peak will do the job.
UAE GDP per capita from World Bank Open Data. Annual, series since 1975, downloads as a single Excel file. Free.
The per capita figure matters more than total GDP because UAE population went from roughly 1.6 million in 2008 to 3.8 million in 2026. Total GDP would mask the actual price-to-income relationship. Use per capita for any country where population has grown materially through the cycle.
You can also connect a market data MCP if you want the macro series pulling automatically. Settings, add MCP, no code required. The Dubai run on this page used two file uploads, not MCP. Both routes work. File uploads are simpler for a first run.
Drop both files in. Paste the locked prompt below. Codex returns clarifying questions and a structured analysis plan before it touches the data. Read the plan, refine if needed, then tell it to go.
I am analyzing the changes of UAE GDP per capita in comparison to Dubai house prices. Specifically, I would like to understand periods of property price acceleration relative to per capita economic growth and look at pricing in Dubai today in comparison to the 2007-2008 peak period. I am attaching historical Dubai property price data and UAE GDP per capita data via excel. Please ask clarifying questions.
A plan to align the monthly price series with the annual income series, then compute price growth versus income growth over rolling windows. A separate compare-to-2008-peak ratio for current pricing. A list of divergence periods, ranked by gap size.
Answer the clarifying questions before approving the plan. They usually cover currency basis, inflation adjustment, and which year to anchor the 2008 peak comparison against. Once those are locked, tell it to proceed.
The price series will have gaps, outlier months, and at least one methodology change. Codex flags each issue, proposes a fix, and waits. Nothing gets cleaned without your approval. That protects the fidelity of the analysis the whole way through.
Missing months in the early years. Reporting cadence on the index was lighter in 2008 and 2009. Codex proposed interpolating between known points and flagged the segments where the interpolation carries the most uncertainty.
Outlier price points around launch events. A handful of months sit well above or below the trend line, usually tied to a large new launch skewing the average. Codex proposed winsorising those months rather than dropping them outright.
A methodology change in the index mid-series. The way the index was computed shifted at one point. Codex proposed rebasing the post-change segment back onto the pre-change basis so the long-run comparison holds.
Codex shows the issue, proposes the fix, and pauses for a yes or no. You can also ask for the alternative approach before approving. Each decision is logged so the cleaned dataset is auditable end to end.
The slow part of the run is reading what Codex is asking. Take the time. This is where the analysis stays honest.
The methodology change is the most load-bearing call in the cleaning step. If the change rebases the series or shifts the basket, pre-change and post-change numbers are not directly comparable. Codex's proposal will usually be to align them. Confirm that is the right call for your specific index before approving, against the index provider's own notes.
Once the dataset is clean, Codex produces a written insights document and a succinct presentation off the same numbers. Both stay live against the cleaned dataset, so re-running with a fresh price file regenerates them in seconds.
Dubai property prices were 36.9% above the 2008 peak by May 2026. UAE per capita GDP was only 0.4% higher over the same window. The two series move together until around 2013. From 2013 onwards, prices run ahead of incomes for sustained periods.
The strongest divergence years were 2013, 2014, 2023, and 2024. In each of those years, price growth outpaced per capita income growth by a wide margin. 2013 and 2014 sat at the top of a cycle. 2023 and 2024 is the more recent stretch.
Prices have risen far faster than incomes. The implication is selectivity. You cannot rely on buying any unit and expecting the market to do the heavy lifting.
The opportunity now is in identifying assets with durable rental demand and a reason to keep attracting tenants and buyers over the long term. Location, building quality, service charge structure, and tenant mix matter more in a stretched cycle than in a rising one.
Codex is pattern matching off two indices. The pattern is real. The conclusion you draw from it is still yours. Pressure-test the divergence years against what you know was happening in the market. If the dates match your own recall, the analysis is anchored. If they do not, push back and ask Codex to recheck the segment.
Codex runs inside ChatGPT. The per capita series is free from World Bank Open Data. The price index is the one piece that may carry a cost, depending on your market.
A paid ChatGPT plan with Codex access is the only software spend. The free tier does not include Codex. Check what your current plan allows before you start.
From a blank Codex task to a cleaned dataset, an insights document, and a presentation. Most of the time is reading the data-issue approvals as they come up, not waiting on compute.
UAE GDP per capita is a free Excel download from data.worldbank.org. Same for every country on the portal. No login, no API key.
You sit down on Monday. The latest monthly price file lands. Codex re-runs the analysis against the new month and updates the insights document and the presentation. You glance, sense-check the divergence chart, and ask the question only you can ask.
Eighteen years of messy index data, condensed into one comparison your market lives or dies by. That is the trade Day 18 ships.
Every prompt in this 30-day series is one capability. FourthspaceOS bundles all of them into a single product: underwriting, comps, market research, deal sourcing, portfolio tracking, and investor reporting. The product runs natively on the same Anthropic agents and plugins you are learning to use this month.
Waitlist members get founding pricing, early access, and one new prompt delivered each day for the next 30 days.
Founding pricing locks in for waitlist members.
One email a day for 30 days. No spam. Unsubscribe whenever.
Send them this page. The price index and the per capita income series are both public. The analysis is free. The exposure to a 2008-style stretch in any market is something every buyer should be able to chart for themselves.
Find me on Instagram ↗Grab a monthly price index and the World Bank per capita series for your country. Upload both into a fresh Codex task. Paste the prompt from step 02 and answer the clarifying questions. Ten minutes from upload to a cleaned dataset, an insights document, and a presentation.
Start at step 01 ↑