Looking to invest in real estate? Here's how data can help.

As the market continues to heat up, real estate investors are looking for ways to capitalize on the opportunities. While some are focused on flipping properties, others are taking a more long-term view and investing in rental properties. The demand for rentals remains high as home prices continue to rise and more people are priced out of the housing market. This is especially true for millennials who are now the largest group of renters. Interest rates are still low by historical standards which makes this still a good time to buy property, especially because new tax laws allow investors to deduct up to 20% of their rental income.
Data can help make smart real estate investment decisions in this environment. There are a number of online tools that provide detailed information on rental markets, including vacancy rates, average rents, and expected rent growth. The rise of digital technologies and the sophistication of processing tools is fueling a boom in the volume and diversity of data that is now available to generate powerful insights to drive results.
For starters, even a simple data source such as Google Trends can help generate useful insights. Users can tap into the unofficial Google Trends API called Pytrends to access and extract information using a simple interface. Then, they can use simple search terms to access useful and powerful data.
Powerful Insights from a Single Search Term
Even single search terms can generate meaningful insights that can be used to drive investment decisions. For example, a simple search term "home for sale" and a location setting to a single state like Texas in Pytrends produced the following time-series plot.

Even this simple time series data is quite useful. You can easily average the data for every three months and produce a quarterly dataset. You can then join the data with the House Price Index (HPI) data that is readily available from the Federal Housing Finance Agency. These simple operations can help understand the relationship between the price index and the number of homes for sale, which could be a proxy for housing demand.


Plotting the two pieces of data can also help generate graphs like the one shown above which can provide additional insights. The plot shows that even though the demand for houses follows a seasonal pattern, its overall trend seems to follow the HPI. The correlations between the demand data and the seasonally adjusted and non-adjusted HPI can also be computed. In this case, it turns out to be 72.4% and 73.6% over the 11-year period from 2010 through 2021. You can also smoothen the interest index or remove seasonality to check if it produces better results.
Comparing Markets in Different Regions
Comparing the interest in housing in different regions is also another way to evaluate the market. It is easy to compare and contrast and display the information in a user-friendly map. You can extract data at different levels of granularity using Pytrends and display it using Plotly. This simple combination can show the interest in housing in different states or metropolitan areas over a specified period.

Deeper Insights Using Multiple Search Terms
Using multiple related terms or keywords and a specified location in Pytrends can provide even deeper insights. Using five keywords like "home for sale," "property for sale," "house for sale," "homes for sale," and "houses for sales” and setting the location to Texas produced the following graph.

This graph shows that "homes for sale" is clearly the most popular keyword among the five but that people use other keywords as well. You can use this additional information to augment and refine the analyses conducted earlier with a single keyword to build better proxies for demand, compare them with HPI and optimize search advertising.
Insights from Word Combinations
Pairing search terms can generate very different insights. There is a very useful indicator called the Google U.S. Housing Market BUSE Index, ("Trend-Spotting in The Housing Market") that is defined as the ratio of the buying and selling Google searches in the U.S. real estate category. Businesses and policymakers can use this ratio as an early warning indicator as well.
Users can construct this index using two single search term datasets, one for the "buy -sell" keyword and another for the "sell -buy" keyword. The "buy -sell" keyword generates results for searches that contain the word "buy," but excludes those that include the word "sell." The combination can then provide the data for constructing the BUSE Index for a specific location, such as the New York metropolitan area. The data can then be smoothened to cancel seasonal variation and noise (Details on developing the index can be found here).
Then, a plot of BUSE with the S&P/Case-Shiller Home Price Indices for say the New York metropolitan area can provide additional insights such as showing the inverse relationship between the two. As the plot shows, as the price index rises, the buy-to-sell ratio falls.

As these examples show, even simple data, such as Google Trends, can provide interesting and powerful insights into the status of specific real estate markets. However, the power of data increases dramatically when you combine it with other data that provide additional, unique signals. Businesses in the real estate industry and policymakers as well as individual investors in the real estate can use such combinations of datasets to drive business or personal investment results or design interventions to reach policy goals.