Data science offers CRE the ability to act in sync with rapidly changing markets
![placeholder](/_next/image/?url=https%3A%2F%2Fimages.ctfassets.net%2F8jgyidtgyr4v%2F6vGEYY2A18oq4LBN2I0KNN%2F123f2629ca9cebeab15813f843068ee8%2FInsight_69_-_The_Importance_Of_Data_Science_In_CRE_Industry.jpg&w=3840&q=75)
Key highlights
Data science offers rewarding opportunities to answer questions and prepare hypotheses almost in real-time
Data science integrates alternative data sets, machine learning technologies, and predictive analytics to extract new knowledge from data and amplify the ability to forecast and nowcast, which empowers CRE leaders to act quickly and astutely as markets shift
Given that real estate investments are valued on current and future cash flows, using alternative data to produce nowcasts and forecasts ensures that decisions are based on current assumptions rather than lagging indicators
Applying advanced analytics – statistical analysis, modelling, data mining, scoring and machine learning – to these data sets delivers a more accurate perspective on what is happening today (nowcast) and what is likely to happen in the future (forecast)
In a challenging macro-economic environment, real estate investors, owners, and managers continue to face new challenges and time pressures in determining the right steps to take next
When market uncertainty runs high – real estate investors, owners, and managers face challenges and time pressures in determining the right steps to take next. Fortunately, data science offers rewarding opportunities to answer questions and prepare hypotheses almost in real-time, which previously may have required weeks or months.
Data science integrates alternative data sets, machine learning technologies, and predictive analytics to extract new knowledge from data and amplify the ability to forecast and nowcast. Bottom line: this offers CRE leaders the ability to act quickly and astutely as markets shift.
Nowcasts, for example, are predictions based on high-frequency data, like mobile or retail transactions. These enable us to understand what happened very recently, what is happening currently, and what will happen in the near future.
Say, for example, a real-estate fund wants to expand a portfolio of multi-family buildings. A nowcast can provide clear and contemporaneous market insights by utilizing machine learning within a massive collection of neighborhood-level, near- or real-time data. Machine learning algorithms can rapidly combine both macro and hyper-local forecasts to prioritize cities and neighborhoods with the highest demand for multi-family housing. This would enable the fund to identify buildings in areas that are currently undervalued but rising in demand.
Given that real estate investments are valued on current and future cash flows, using alternative data to produce nowcasts and forecasts ensures that decisions are based on current assumptions rather than lagging indicators.
This offers real estate investors and organizations significant strategic advantages. Combining conventional demographic and migration data with near real-time alternative data in advanced data models can discern significant trends well before traditional statistical metrics. They can identify markets with the greatest potential, assess growth or demand for specific asset classes or sub-classes in particular markets, and accurately calculate the future value of potential acquisitions.
And, applying advanced analytics – statistical analysis, modelling, data mining, scoring and machine learning – to these data sets delivers a more accurate perspective on what is happening today (nowcast) and what is likely to happen in the future (forecast).
Using this approach can optimize real estate investment and acquisition strategies, uncovering hidden opportunities and risks.
Data science
Uses scientific processes, algorithms, and models to extract actionable insights from raw data.
Alternative data
Is data from non-traditional sources, such as mobility and location data, retail spending activity, real estate listings and sales, and web activity.
Predictive analytics
Is a data science process that identifies the likelihood of future outcomes based on historical data.
Case in point: Identify multi-family investment opportunities by looking at migration trends
Here is a good example of how one real estate fund utilized nowcasting to acquire insights that were never previously available.
To inform their portfolio strategy and to demonstrate to investors they had a good understanding of its markets, opportunities, and risks, this fund needed a nowcast of migration trends for a key market in which it had significant exposure — Las Vegas.
Interestingly, during the pandemic, multi-family rents in the Las Vegas market were stable or increasing while the employment situation was among the worst in the country. Fund managers hypothesized there was an influx of people moving to Las Vegas but had no evidence to support this hypothesis. Estimates based on traditional data sources like the US Census Bureau were too outdated to be useful.
To address this need, we integrated mobile phone GPS data with neighborhood-level household characteristic data on a machine learning platform, which allowed us to nowcast household movement patterns and overall market health at the neighborhood level.
From there, we could identify several key factors at play. First, a significant number of new Las Vegas inhabitants were arriving from southern California – Anaheim, Los Angeles and Riverside in particular. Second, incoming residents had, on average, household incomes that were $20,000 higher than those of current Las Vegas residents. For these new arrivals, Las Vegas rental rates were very affordable. As a result, occupancy and rental rates through the pandemic remained healthier in Las Vegas as compared with other markets.
Case in point: Find the markets where the return-to-office is most advanced
For another project, our client’s goal was to assess foot traffic recovery patterns within major office neighbourhoods across the US by comparing differences between 2019 and the current quarter (Q3 2021). By combining estimates of office-based employment, and mobile phone-based foot traffic within a machine-learning platform, we were able to benchmark current traffic volume with that of 2019.
This analysis showed, as an example, that Atlanta had one of the strongest recoveries of foot traffic within its office neighborhoods. San Francisco, by comparison, was still about 50% below 2019 levels. Not surprisingly, in 2021 Atlanta also had one of the strongest-performing office markets, while San Francisco lagged the nation.
Many uses for advanced data science in real estate
These cases demonstrate how advanced data science can accurately nowcast and forecast shifting demand for real estate, enabling CRE investors and organizations to better understand the factors impacting performance. Here are more examples.
Utilize location, demand shifts and physical attributes to identify why certain assets experience stronger net operating impact (NOI) performance than others
Target acquisition opportunities with exceptional potential by leveraging machine learning to forecast how certain properties can be expected to perform
Use alternative market data and machine learning model predictions to identify assets that can be expected to underperform
When it comes to decision-making in the real estate industry today, the pandemic demonstrated how quickly markets can change. Data science gives CRE investors the advantage of real-time insights to quickly calibrate strategy and make decisions with confidence and speed.
Author
![undefined's Profile](/_next/image/?url=https%3A%2F%2Fimages.ctfassets.net%2F8jgyidtgyr4v%2F6IneUNcntX8W3ng262nK9b%2Fb60b8689106a9c4b2acd1ee2f9c1b487%2FMichael_Clawar.png&w=3840&q=75)
Michael Clawar
VP, Head of Data & Analytics
Author
![undefined's Profile](/_next/image/?url=https%3A%2F%2Fimages.ctfassets.net%2F8jgyidtgyr4v%2F6IneUNcntX8W3ng262nK9b%2Fb60b8689106a9c4b2acd1ee2f9c1b487%2FMichael_Clawar.png&w=3840&q=75)
Michael Clawar
VP, Head of Data & Analytics