How can we improve the design, monitoring and evaluation of temporary changes to public realm spaces? Can novel uses of urban data help us do this better?
These were the questions we first asked during a conversation between the Edinburgh Futures Institute and Edinburgh Living Lab at the University of Edinburgh, Jacobs, the City of Edinburgh and Sustrans in late 2019. This conversation evolved into a project around the Open Streets programme in Edinburgh, and subsequently the Spaces for People temporary street interventions. Both project phases explored how urban data could potentially provide valuable insights to inform city decision-making.
The first phase — in early-mid 2020 — resulted in a resource and report aiming to support data-informed and community-centred decision-making for the public realm. This focussed on digital engagement methods.
It included a framework for understanding different stages of a public space transformation process and identifying key types of data that could inform decision-making throughout that process. Case studies illustrated what this could look like in different cities. You can view the Phase 1 report here.
The second phase of this work took place in late 2020. It aimed to pilot and test two of these novel methods of harnessing otherwise ‘unseen’ datasets in practice, to support monitoring and evaluation of live changes to the public realm.
Specifically, this analysis focussed on the temporary interventions installed via the Spaces for People scheme in Edinburgh. This resulted in the ‘Piloting Novel M&E Methods for Street Interventions’ report, also viewable as a 3 page summary. Whilst a small-scale initial study, this yielded some interesting findings about both the potential and limitations of these methods.
The two methods tested were:
- Sentiment analysis of Edinburgh-wide Twitter data using #SpacesForPeople.
- Spatial data analysis of Just Eat cycle hire data (thanks to provision of a large dataset by Serco) to help reveal changes in cyclist numbers and chosen routes before and after the public realm intervention.
The piece of work also tested out how combining these novel technology-based data-driven methods with qualitative data and other more traditional data collection methods, such as interviews or surveys, allows the triangulation of findings that provide more robust insights.
Sentiment analysis findings
One key finding of the study was that sentiment analysis is best suited to situations where changes to the public realm can produce large and easily traced datasets. An example of this is city-wide schemes such as #SpacesForPeople with a dedicated social-media hashtag (that can then be refined by location) or particular keywords.
Our study found that tweets from Edinburgh using #SpacesForPeople were typically slightly positive — 0 to +0.3 on a scale of -1 to 1. By using keywords that either included street names or particular design infrastructure elements (e.g. ‘bollards’ or ‘cycle lane’), we could start to understand how different streets’ interventions within the city-wide scheme or particular design elements were faring in terms of public support.
Sentiment analysis is useful in these situations to gather broad-brush indications of the level of positive/neutral/negative support, with other complementary methods providing further insight. Use of sentiment analysis via a live-updating tool (such the one below trialled for this project) could also potentially be used as an early warning system flagging any spikes in positive/negative sentiment — tracking in real-time and flagging possible issues needing further investigation in a timely way.
Findings from GPS trace cycle hire data
Analysis of the Just Eat cycle hire GPS trace data also revealed useful and otherwise unseen insights about how changes to the street environment affect cyclist numbers and routes chosen.
Using a series of ‘gates’ that isolated data for the street of interest, the team could analyse GPS traces for cyclist movement to start to understand the ‘success’ of that intervention in terms of how well it was used by cyclists. This was done by using indicators such as the ratio of cyclists choosing to use this route vs viable alternatives, as well as cyclist numbers both pre and post implementation.
For the Spaces for People street changes implemented on George IV Bridge, this method revealed a 25% increase in cyclist numbers on this route immediately following the Spaces for People changes to the public realm. Compared to overall city-level cyclist numbers, which remained static, this suggested that the changes were having a positive influence on cyclists using George IV Bridge . Further qualitative data would help to understand the motivations and reasons behind this in more detail.
So, can new ways of gathering data help improve monitoring and evaluation of changes to our public realm?
I would say yes, based on the methods piloted in this project. However, it is the combination and triangulation of insights and feedback from multiple data sources (qualitative and quantitative, real-time and static) that can provide the most robust and holistic understanding of how the public realm, and any changes to the street environment, are performing.
Whilst novel data-driven techniques for the public realm, such as use of sentiment analysis and cycle hire spatial datasets, can add immense value and insight, it is vital to combine these with in-person and qualitative approaches to give the full picture and understand the ‘why’ behind the numbers.
This initial pilot study showed the potential for different data sources to be combined to support policy and operational decisions for the city, allowing a more structured feedback loop between changes to the public realm on the ground and city-level decision-making.
Perhaps it is finding a more structured way for this feedback loop to become embedded in everyday city decision-making that would be the biggest step to improving the design, monitoring and evaluation of our public realm?
This work was completed by the Edinburgh Futures Institute, via their Smart Places workstream, in collaboration with the Data Driven Innovation initiative at the University of Edinburgh. It follows on from work with CEC and Jacobs around Open Streets.
The project was delivered via a small ESRC Impact Accelerator fund, which allowed close collaboration with Jacobs as industry partner on this project, with kind support in-kind from CEC Officers (via provision of dates Spaces for People interventions/infrastructure were introduced) and Serco (providing a large Just Eat dataset).