Applications of Data Science for Architecture & Real Estate
With 5.6 million commercial buildings in the US (source: US Energy Information Administration), there are plentiful opportunities for data collection and application in the built envionment. Looking at building data holistically draws information from architecture, engineering, construction and real estate. Why does building information matter, and what can be done with it?
Design Data
Design data includes all of the information architects and engineers use to come up with the building design. This information is codified on construction documents, typically presented in pdf form to the local building department for the purpose of obtaining a permit. For the uninitiated, construction documents were called “blueprints” back in the day.
However, design data has many other uses other than permitting. It can be leveraged to analyze the building, determine its quality, compare it to other buildings, and identify opportunities for renovation or expansion.
Some building departments such as NYC and San Franscisco publicly share data collected during the permitting process. NYC Open Data shares building footprint shape files. The permit itself also contains data — such as the type of project, date, address and a description of the work.
To a casual observer this information seems rather dry. Consider, however:
- Building permits are a strong indicator of economic activity
- Construction activity precedes real estate activity
- Market trends are constantly changing with demographic and economic conditions
All of this means that the applications for predictive analytics are enormous.
BIM Data
BIM, known as Building Information Modeling, is the predominant method of designing and documenting buildings for construction today. BIM data is a subset of design data. BIM software such as Revit and Vectorworks consist of a 3D model embedded with information about itself. A variety of information can be obtained from BIM models such as the materials used, building code information, and detailed quantitative measurement of the building size and shape. BIM data can be used for detailed quantitative analysis of the building.
Why is BIM data important and how can it be leveraged? At the end of the design process, the BIM model contains nearly all the information that goes on the construction documents — a roadmap for producing a building. With the industry moving towards prefabrication and automation (consider Katerra and Project Frog) BIM data can be leveraged to build buildings faster and more accurately, saving time and money.
Construction Data
Construction data includes information such as construction cost, time, number of bidders and bids, whether any accidents occurred, the type of contract and project delivery method; i.e. whether it was design-build, traditional design-bid-build, and whether a construction manager was utilized. Dodge Data & Analytics and Construct Connect are two industry leaders tracking construction data. Construction data can be used for market research, cost analysis and improving the efficiency of the construction process.
Why does construction data matter? Slowing construction starts are a well-known predictor of recessions. Increasing construction costs may correlate with high corporate debt rates, the rising cost of secondary education, and municipal budget deficits. Studying the relationship of construction activity to any one of these relationships from a data science perspective could yield tremendous insight.
Real Estate Data
Real Estate Data includes occupancy data, tenant data, financial information about the building, and a snapshot of basic building information like gross square feet, number of floors, and leasable area. Co-Star is the industry leading provider of general real estate data. The main purpose of this information is to help brokers close deals; it can also be used for property comparison and appraisal.
Facilities & Operational Data
Facilities and Operations data includes information tracked by the building owner as a part of their process of using the building. This can include things like the cost of repairs, energy usage, revenue, profit, utility accounts and user data.
Architects sometimes provide services for Post-Occupancy Evaluation (POE) which involves determining how well the building suits the needs of the users through a series of questions and analysis. Part of WeWork’s rapid rise was driven by the use of digitized POE’s. The digitized POE’s enabled users to provide instantaneous feedback on problems in the building to the building management, resulting in faster repair times and happier occupants.
Risk Data
Risk data about buildings includes the potential for fire, collapse, structural deterioration, flooding, damage from an earthquake, or other hazards. While some of this information is held by the insurance industry, a lot of it is not typically codified into premiums.
Other risk data is out of date. For example, the FEMA flood maps are considered unreliable if more than 5 years have passed since an update, and roughly 2/3 of the FEMA maps are older than 5 years. Obtaining accurate risk data is challenging, however the benefits to the wellbeing of the built environment is well worth the cost. For example, knowing the accurate flood risk prior to construction would enable designers to design built-in measures to allow the building to withstand flooding. Without this data, the potential for damage dramatically increases.
Final Thoughts
Looking at an overview of the possible data that can be analyzed for buildings, there is rich opportunity for startups in the building industry data space as well as the potential for massive impact on humanity by more effectively leveraging building data.