Data analytics is booming, and real estate professionals are starting to catch up.
Predictive analytics is becoming an omnipresent force in various business sectors to power decision making and strategizing processes. According to the KPMG Global PropTech Survey in October 2019, 58% of real estate companies had a digital strategy. 59% of the development firms believe that they will invest in IT, digital or PropTech collaboration in 2020, and 78% believe that they will do so in the next five years. Industry professionals now see the need to leverage technology to give them a competitive edge.
Real estate professionals today are in a position to drive more value than ever by leveraging data analytics and machine learning. “While the real estate and property industries may have been slower to adapt to changing technologies, they are catching up, and AI and machine learning will help make the data they are already collecting more actionable,” — Bernard Marr, an internationally bestselling author on topics related to technology and innovation.
How can you make the most of these changes though?
We’ve come to realize that combining qualitative variables and indicators with traditional factors such as household income and vacancy lead to extremely powerful insights. A McKinsey report on real estate data transformation estimates the predictive power of non-traditional factors (e.g. property’s proximity to a 4-star hotel or number of cafes within a mile) drives 58% of a model’s predictive power
However, the idea of “non-traditional predictive factors” or even “big data” in the context of real estate development is difficult to grasp. We believe in openness and transparency at Ascend and want to communicate how these abstract concepts factor into our data-driven solutions.
The rest of this article features a preliminary geospatial analysis on the impact of neighborhood trends on property price and transaction frequency in the city of Vancouver, British Columbia. This analysis is not necessarily representative of our products in terms of scope, variable choices and functionalities. It is meant to showcase that though qualitative data can be ambiguous, they have the potential to generate insights and lead to value creation for developers and brokers.
We know less crime leads to higher property prices, but few analyses have been done in the Metro Vancouver Area.
Less crime = higher prices is a logical conclusion — but current valuations only show trends on a neighbourhood level. A large body of research has been conducted on the impact of crime on price transactions in US and European Markets (The impact of crime on property values: Research roundup by the Harvard Kennedy School). However, few analyses have been done on the Vancouver region or have incorporated geo-spatial visualizations.
We pulled crime data from 2013 to 2018 (mischief, break and entry, theft, offenses against individuals) to visualise crime rate change over time. Our goal is to highlight whether higher housing prices in a region are associated with a lower crime rate in the same region. Certain crimes are excluded for privacy and investigative reasons.
We pulled data from 2018 (shown above) and observed a reverse correlation between crime rate and housing price after pulling data from 2018 as shown above.What does this mean?
Housing prices increase when crime decreases. Intuitive right? The analytics behind finding this data is harder. We utilized over 620,000 data points to layer a year-on-year trend of micro-zones within Vancouver. Comparing the two maps — you can see regions with hollow spots for crime (left map) such as the north of the Kerrisdale neighborhood, Oakridge and along the southern portion of Kingsway. In these areas, housing prices were higher than in other parts of the city.
Some areas are less clear cut than others. The downtown region and north coast of Vancouver experience the highest crime rate, though these regions are known for high housing prices. There’s a possible explanation for this: these are the more affluent areas of the city. We suspect that the high crime rate is mainly represented by crimes such as theft and break-and-enter — types of crime which tend to happen in wealthier regions of cities. 
The above time-series heat-map shows how crime rate changes over the time period 2013–2018. Crime rates decrease overtime along Kingsway in the Kensington area and also towards the north in the Hasting-Sunrise neighborhood.
These two areas also saw large jumps of housing price increase from 2017 to 2018, which supports a connection between a decrease trend in crime rate and an increase in housing price.
This data leads to several thoughts that can be helpful to developers. First, the known sentiment that people are more willing to pay for houses situated in areas that are expected to be safer in the future. Second, areas with a higher percentage of young people can be related to both a lower crime rate and potential housing price growth. This is seen in the Kensington and Hasting-Sunrise areas, which are among some of the most racially diverse and young areas of the city. Third, improving areas are correlated with the presence of active business associations and community development efforts. A well-managed, racially diverse community could be the key to a declining crime rate and the prosperity of its real estate market.
The best Yelp businesses catalyze booming housing markets.
We were curious to understand how the quality of surrounding businesses impacts housing prices of different neighborhoods in Vancouver (inspired by an HBS study that highlights how opening a Starbucks in an area drives housing prices up by 0.5% within a year). We utilized Yelp data on customer rating and business location coordinates to generate the following heat-map.
Within Vancouver, high-quality businesses are concentrated in Downtown, Kitsilano, central Kerrisdale, and along Kingsway. First, we take a look at how the geospatial heat-map of the business rating compares to the housing price heat-map. Immediately, we notice that the impact on housing prices of a business center radiates into its surrounding neighborhood. Yet, the business centers themselves seem to have a lower housing price. Why?
We think the following:
1) Higher quality businesses attract housing demand, and this high demand drives the housing price up.
2) However, because within the core business areas, housing transactions seem to be low, we suspect that residents are less willing to live in an area where businesses are very concentrated.
3) Another interpretation of this observation could be that since there are more businesses in the area, there are less residential housing that could be transacted in the area.
Ascend is uniquely positioned to use qualitative variables and indicators, alongside other traditional factors such as household income and vacancy, as powerful tools to generate insights for developers and brokers. We’re here to help you find better properties, faster. We’d love to speak with you at: email@example.com
Ascend is an AI platform that synthesizes zoning information and development activity data points across cities to suggest arbitrage and value creation opportunities. We analyze nontraditional (e.g. sentiment analysis, housing improvement factors) and traditional (e.g. zoning, rezoning potential, construction activity level, property’s management profile) data to help real estate developers, brokers and city governments uncover undervalued development opportunities.
* Our main crime dataset includes mischief, break and enter residential, other theft, and offense against a person, since we are limited to open source crime data published by the city police department.