Introduction to Sustainability Metrics in Real Estate

Understanding sustainability metrics like electricity and water usage, and CO2 emissions is crucial for real estate management. However, interpreting these metrics can be challenging without a proper context. In our partnership with Varig, we aim to transform how these metrics are perceived and used in day-to-day real estate operations.

The Power of Comparative Analysis

To make sustainability metrics more comprehensible, we adopted a comparison approach. This method allows clients to understand their consumption in relation to similar peers. For instance, knowing whether 10 kg of CO2 is a high emission is clearer when compared to peers emitting an average of 5 kg. This comparison not only aids in understanding but also gradually builds familiarity with what these metrics signify.

Building Relevant Peer Groups with Machine Learning

Creating relevant comparisons involves grouping similar assets, which could be challenging when comparing diverse properties like warehouses and hospitals. We tackled this by developing an algorithm powered by machine learning, which considers various asset features such as building size and usage type. The algorithm helps identify the most relevant peer group by analyzing the importance of each feature, ensuring the comparisons are apt and valuable.

Addressing Privacy Concerns in Data Sharing

In presenting these comparisons, we ensure the privacy of client data. Instead of individual metrics, we provide averages from larger peer groups. This approach protects individual data while still offering valuable insights through benchmarking.

Technical Implementation: Data Collection and Analysis

Our process starts with a detailed collection and analysis of building data. By plotting and examining the distribution of building features and energy consumption, we identify and exclude anomalies to refine our peer group. The use of visual tools like distribution plots and heatmaps helps in understanding correlations between building features and energy usage.

Peer Selection Algorithm and Benchmark Presentation

The peer selection algorithm we developed sequentially adjusts the set of considered features based on their importance, optimizing the search for a sufficiently large and relevant peer group. We then present the findings through a peer grouping quality metric and a comparative performance metric, enabling clients to assess their position against benchmarks and make informed decisions.

Conclusion: Impact of Data-Driven Sustainability in Real Estate

By leveraging machine learning and data-driven insights, we help real estate owners not just understand but also act on sustainability metrics. This approach not only aids in immediate operational adjustments but also supports long-term sustainability goals in the real estate industry.