My Journey Through Econometrics and Urban Design

 

How do you measure the 'soul' of a city? As an architect and urban designer, I've often struggled with conveying the subtle, intangible benefits of urban design. Elements like "quality of space" or "sense of place" are central to creating vibrant urban environments, yet they resist easy quantification. In our increasingly data-driven world, this presents a challenge. Numbers are the language of policymakers and stakeholders, but how do we translate the human experience of a city into data?

After sharing on LinkedIn that I completed the "Econometrics: Methods and Applications" course from Erasmus University Rotterdam, several friends reached out, curious about how econometrics fits into my work. This blog is a response to them and anyone interested in the intersection of design and data.

Aspects like improved pedestrian connectivity, enhanced social interactions, etc are central to urban placemaking but don't easily fit into spreadsheets or reports. This led me to wonder: how do economists manage to quantify complex, intangible concepts and gain widespread acceptance? The answer I found was econometrics.

Determined to bridge the gap between the qualitative aspects of urban design and quantitative data, I enrolled in that econometrics course. Initially, I wasn't sure how mathematical models would fit into my work. The world of equations, regressions, and time series felt daunting—far more math-heavy than anything I'd tackled before. But as I progressed, I discovered how incredibly rewarding and applicable these tools could be.

The course covered everything from simple and multiple regression analysis to hypothesis testing, model diagnostics, time series analysis, and panel data techniques. The hands-on approach of the course meant I got to use statistical software to estimate models, interpret regression results, and tackle common issues like multicollinearity and heteroskedasticity. This journey not only sharpened my analytical skills but also empowered me to critically evaluate empirical studies, carry out my own econometric analyses, and apply quantitative methods to assess real-world problems.

Ordinary Least Squares method: This is one of the most fundamental tools in econometrics. It’s used for estimating the relationships between variables. Say, you want to understand how the size of a public square impacts the number of people using it. Starting with Simple Regression models, one can work up to Multiple Regressions and that helps you quantify that relationship by looking at the data you’ve collected and drawing a “best fit” line through it. It can also look into how things like increased pedestrian connectivity can improve the vitality of a space.

For instance, one project involved analysing a dataset on housing prices to determine what factors most significantly affect property values. Using multiple regression analysis, I examined variables like location, square footage, number of bedrooms, and proximity to amenities. This exercise required cleaning the data, selecting the appropriate model, and interpreting the results to draw meaningful conclusions.

Binary Outcome Evaluation: Binary outcome models fascinated me, especially their potential to answer yes/no questions. While commonly used in finance to predict consumer behaviour, I see potential applications in urban design, like determining whether introducing informal street vendors increases foot traffic and economic activity.

In the context of urban design, binary outcomes can extend to various factors like determining whether better street lighting influences people’s feeling of safety, or if incorporating more green space encourages more public usage. It turns subjective feelings into tangible outcomes that you can work with, and this changes how you approach the design itself. By generating Odds-ratio and Maximum Likelihood, one can better understand outcome probabilities.

Time Series: Time series analysis was another monster altogether. Slightly steeper learning curve too. It allows to account for change over time. The training was for forecasting Industrial Production based on a Composite Leading Index. Running tests to determine Stationarity, transforming the data to achieve Stationarity, trying out AR and ADL models (that’s Auto Regressive and Augmented Dickey Fuller for those who are curious enough!), checking for Granger Causality, to finally predicting the forecasts, and further evaluating both models to select the better one.

Cities are dynamic, and people's behaviours shift over time. This tool allows one to track changes. This could include pedestrian movement patterns, traffic congestion, or changes in land use. Further, with remote sensing data of a place spread over time can be an interesting potential application.

One application using time series to analyse how different design interventions affect urban flow over the long term. It could also provide clues on user charges to be levied for financial feasibility assessment, or may be use data from sensors on movement patterns and create predicted models.  With time series, it is possible to analyse patterns that emerge slowly, providing solid evidence for or against certain design interventions.

The Role of R software:  Learning R, an open-source statistical software, was like learning a new language. It wasn't plug-and-play, but the power it offers is unmatched. I remember spending hours troubleshooting code, but the satisfaction of seeing my data visualizations come to life was worth every minute. R allows you to run sophisticated econometric models, visualize data, and make predictions all in one place. I used the R Studio version.

What surprised me was how quickly I started to see connections between the mathematical models and the real-world urban issues. Whether you're an architect, engineer, or even a social scientist, the ability to handle data in R is becoming essential. Trust me, if I can learn it, so can you.

I've realized that quantitative analysis doesn't replace the creativity or intuition inherent in design—it enhances it. By justifying decisions with data, we can better communicate the value of our work to stakeholders who rely on numbers, like city planners or policymakers. It strengthens our position, ensuring that design-driven decisions aren't easily swayed by subjective opinions.

So basically, econometrics gives one the tools to talk about urban design in a language that everyone understands—numbers. It doesn’t diminish the beauty or the human aspect of the work we do as architects or urban designers. If anything, it strengthens it. We’re now able to back up our claims with data, making it harder for subjective opinions to sway what should be objective, design-driven decisions.

Learning Econometrics has taught me that the gap between subjective design quality and quantitative analysis isn't as wide as I once thought. With the right tools, it's possible to bridge that gap, giving urban design the rigor and respect it deserves in data-driven discussions.

Now, I'm venturing into learning GIS (Geographic Information Systems) software, and I find that my newfound data skills are easing this journey. I'm excited to combine econometrics and GIS to further explore urban design from a data-driven perspective. GIS allows for spatial analysis, which, when combined with econometric models, can provide deeper insights into urban patterns and relationships.