Connectivity & Data
Governance and Citizen
Energy & Environment
How can cities understand more about all of their citizens – not just the vocal minority?
There’s a growing understanding that smarter cities will require much greater citizen engagement and involvement. This is the only way to ensure that people remain at the centre of city development and that technology is used in line with the stated aim of improving quality of life for residents and visitors.
Action to better include citizens in decision-making is also increasing – albeit more slowly – with more cities beginning to hold consultation exercises such as polls, debates via online platforms, in-person workshops and more.
While these are a step in the right direction, the information gleaned still doesn’t tell the whole story. Typically, a minority of citizens are simply more minded, or able, to participate in these types of activities and share their opinions.
”Everyone has a role in services such as recycling and we need as many opinions as possible to shape services that people want, understand and will respond to.”
To achieve their goal of taking as many citizens’ views into account as possible, cities and other public sector bodies, such as transportation authorities, must go further – they need to reach the ‘silent majority’.
“We want to reach as many people as possible,” commented Mark Saunders, Director of the Ferrovial Services Centre of Excellence for Cities. ”Everyone has a role in services such as recycling and we need as many opinions as possible to shape services that people want, understand and will respond to.”
The good news is that this doesn’t mean badgering citizens for more of their time and data – rather, it’s about listening better. A lot of the useful information cities require is already out there.
Research suggests we generate 2.5 quintillion bytes of data every day. The challenge is making sense of it. Much of this data is untapped, and cities could use it to generate new insights about how their citizens feel, prioritise services, carry out predictive actions and better understand the nuances around specific issues.
Research suggests we generate 2.5 quintillion bytes of data every day.
As David Graham, Chief Innovation Officer, City of Carlsbad, told SmartCitiesWorld in an interview recently: “As members of the public, we already provide so much information about ourselves – voluntarily – to companies, to social media, etc. There’s the opportunity for the city begin to gather and garner some of that information so that we have a more accurate understanding of our residents.
"Additionally, across all the departments, each of your touchpoints with the city is helping to develop your user profile so we have a better understanding of what you want, who you are and the types of services that we can provide.”
One promising area is the advancing field of natural language processing (NLP). NLP leverages artificial neural networks and data science to analyse content from a variety of sources, including social media, web-form information and news feeds, to derive insights and trends. This is all information that cities have access to already and could use better.
NLP automatically extracts aggregated, anonymised opinions, insights and trends from text. These tools can help organisations understand general sentiments such as whether the mood is positive or negative; specifics about service issues – for example, whether a transport service is poorly perceived because it’s late or because the bathrooms aren’t clean; and how services are tracking against known issues such as wait times, bin collections, bill payments and more.
Text and sentiment analysis is increasingly used in the private industrial sector to drive and improve services, using comparable volumes of data to the amount handled by many cities.
Thomas Evans, General Manager of the Ferrovial Services Asset Management team based in Australia, said: “Analytics has helped us considerably in the area of rail asset management where legacy systems are in use.”
“In one project, we found 750,000 free text comments from maintenance technicians – far too many to process, review and derive any insight from. We leveraged free text analysis to develop a heatmap that enabled us to quickly identify the key issues and asset types to focus on.”
Free text differs from surveys because the conversations are unconstrained by set and inherently biased questions. It’s a way for cities to find out what citizens really think and it can shed light on issues that aren’t even on cities’ radars – such as confusing signage causing near-miss accidents, for instance.
Grade.DC.gov uses analytics to improve customer care at participating DC government agencies.
Use of text and sentiment analysis in the public sector is less well documented. Washington DC is one example, with its Grade.DC.Gov platform. Grade.DC.gov uses analytics to improve customer care at participating DC government agencies. Feedback is collected from the website and combined with data from comments posted on social media sites like Twitter. This data allows the DC government to form a grade for each agency.
The Obama administration used sentiment analysis to gauge public opinion to inform policy announcements and campaign messages during the 2012 presidential election campaign. This enabled the team to tweak messages and the strategy accordingly. Trends in sentiment on social media have even been linked with shifts in the stock market.
Carlsbad’s Graham told SmartCitiesWorld that for cities he sees things moving, “towards a CRM (customer/citizen relationship management) approach where you begin to think about your residents like a company thinks about its customers”.
Like most things, text and sentiment analysis must be used wisely and with caveats and appropriate privacy protections in place.
As sophisticated as our technological tools and techniques are, humans remain more complex. Language is used in weird and wonderful ways and is always evolving. People are sarcastic, for example, especially on social media.
Like most things, text and sentiment analysis must be used wisely and with caveats.
If someone tweets: “Train is late again. Brilliant”, or “No bin collection today. Great!” you don’t want this to be counted as a positive sentiment. Tools are advancing to account for these subtleties but text and sentiment analysis must still be combined with other engagement tools and methods to gain a holistic picture.
They could tell you a lot more than you currently know and provide an insight into those crucial things that citizens don’t tell you today.
Saunders said: “Today, collectively, we have the data and we can analyse it, so we should act on it for public good.”
He added: “Engaging and listening to citizens is really important but when we do so, we need to respond to make the services better for all – this includes the vocal minority but especially the silent majority.”
You might also like: