Intelligence in the city
Alain Chiaradia and Louie Sieh of Cardiff University explain the importance of data management for towns and cities…
In the film Her, a man who is in a romantic relationship with his computer’s intelligent operating system (Samantha, voiced by Scarlett Johansson) is appalled to discover that ‘Samantha’ has been chatting up other people – 8,316 of them to be precise. To maintain 8,316 parallel meaningful conversations is impossible for humans because we just do not have the capacity necessary to collect relevant data, the adequate knowledge to convert that data into information, nor the necessary disposition to converse in specific, meaningful and particular ways at 8,316 convenient times.
We are faced with this problem of more data than we can deal with in everyday life; just think of your inbox when you get back from holiday. In the parlance of knowledge management, this problem can be described as an inability to turn data that is located in the world into actionable knowledge which is located in agents, via the deployment of information. Intelligence is the process of making information work for decision making through data identification, collection and analysis, and the deployable knowledge itself.
Local authorities hold large amounts of data which simply remain data and does not become intelligence nor inform action. This data-knowledge gap is a major problem in the effective use of information in the public sector. Management decisions in pursuit of efficiency or economy may exacerbate data-knowledge gaps, even as ever more data is generated. Public administrators need to be alert to the potential of data, to analyse and thus transform data into information, producing intelligence, and to grasp opportunities for acting on that intelligence. Importantly, they need techniques to be able to do this effectively.
In managing knowledge for the administration of towns and cities, leaders would do well to pay attention to the following ways of reimagining of the human-machine interface to produce ‘intelligence’:
First, standard data could be ‘street-based’ rather than ‘area-based’. The way that standard data are formatted can facilitate the automated production of intelligence. With a number of exceptions (e.g. crime data, Ordnance Survey Mastermap and land use, which are street based), currently, most datasets are ‘point data’ or ‘pre-set area data’ (output area, lower super output area, wards, local authority areas). A streetbased, rather than pre-set area data format is more versatile for addressing user-defined and bottom up issues. The anonymised street-based formatting of cencus and related data would have a massive benefit for open data application take up. Street level data formats for all urban management data would dramatically reduce uncertainty about data aggregation into information and thus facilitate the automation of analysis, which in turn produces intelligence. One example is Cardiff University’s spatial Design Network Analysis (sDNA). This tool extracts information from the street network, and the network shape itself becomes the common glue between datasets to establish place rank. This built on Google’s innovation page rank which was to analyse the network itself, and not only the website content.
Second, explore automatic means of transforming data into information. Such means already exist. In the finance sector, this has until recently been the kind of work done by junior analysts, who pull data from terminals, fill in spreadsheets and crunch numbers. This can now be done by machines (e.g. Warren a “virtual market assistant” – like Apple’s Siri, but for investors) that draw inferences, answer questions and recommends action too 1. Machines are taking a bigger role in not just dealing with data, but in the creation and deployment of intelligence. Indeed, ‘agents’, in which knowledge is located, do not have to be human. The agent could be a machine or an organisation combining both human and machine. This rethinking of the human-machine interface could make standard public sector data more available. See for example the Greater Manchester Data Synchronisation Programme.
Third and finally, incorporate ‘imagineering’. The means of converting information into intelligence requires independent thought. If data mining, machine learning and artificial intelligence are extracting patterns and generating information more efficiently and in a more timely manner, ‘human beings’ independent thoughts and imagination should be freed up for imagining what is possible. In an information society, ‘imagineering’ – the explicit evidence-based visioning of “what if” scenarios – will be as necessary to towns and cities as engineering – the explicit mechanics-based imagining of technical possibilities – has been to sophisticated machines. At CardiffUniversity, students of the MA Urban Design do a community involvement-led urban design project in their second semester. In the past 2 years, they have worked together with local community stakeholders to re-imagine Swansea’s historic High Street.
The great challenge is to enable machine intelligence and human imagination to beneficially inform one another. Successful ‘imagineering’ should enable powerful generalisations of data driven intelligence and specificities of human knowledge to interact fruitfully. In towns and cities, this should be in an arena that admits the voices of diverse urban stakeholders. Decision-makers must not succumb to ‘managerialism’ and be frightened of framing their use of data-driven intelligence in the specific knowledge of place. Neither should local stakeholders shy away from framing local issues in the broader context. Data-driven intelligence should not be a prison that inhibits decisions. It should be wielded to set imagination free in pursuit of an optimistic future.
1 FT Technology: rise of the replicants by Richard Waters 03/03/2014
Lecturer in Urban Design
Tel: +44 (0) 29 2087 4640