Why use data?
FACT: too many adult pedestrians are killed, or experience life-changing injuries on Liverpool's roads and, whilst the overall number of collisions is reducing , pedestrian casualties in Liverpool have plateaued since 2008.
So-Mo, have been asked to address this challenge. For this project, we have joined forces with Road Safety Analysis who are providing us with support to analyse the data and will share their not inconsiderable experience in the field of road safety. This blog, written by Tanya Fosdick from RSA explains why a thorough analysis of available data is the best way to bring focus to a complex challenge.
So, what does an exploration of the data involve?
We are going to begin by delving into the STATS19 collision data. STATS19 information is either collected at the scene of a road crash by a police officer or is reported over the counter at a station. It is only for incidents on public roads and where someone has been injured (so doesn’t include damage only crashes or incidents in car parks). The information is not the result of an extensive investigation – it is based on witness information at the time of the crash (whereas collisions resulting in death or life-threatening injury are investigated by specialist collision investigators over a much longer period of time).
So, the dataset does have its limitations. It is thought that there is under-reporting of slight injury collisions (because people don’t always call the police for minor injuries) and that some contributory factors (such as mobile phone use) might not be immediately obvious at the scene. However, STATS19 is a national statistic and is published by the Government annually. It is the best insight available into the extent of injury collisions which occur across the country and contains a wealth of information on how the collision occurred and who was involved (both the casualties and non-injured participants). Whilst it is not perfect, it is the best we have and has been used consistently for collision analysis in Great Britain since 1979.
At Road Safety Analysis, we created MAST Online to allow those in the road safety sector to analyse national STATS19 data in a web-based tool. When undertaking a collision analysis to understand the risks facing adult pedestrians in Liverpool, we don’t start with any assumptions. We systematically work through the STATS19 fields to answer the following questions:
- What? What is the problem they are facing? Are the numbers of adult pedestrians injured in Liverpool typical compared to other areas? Are the numbers of these casualties going up or down?
- When? What is the time of day, day of week and month of year when adult pedestrians are injured? What are the lighting and weather conditions at the time?
- Where? What kind of roads are the pedestrians on? What is the speed limit and is there a junction nearby (and if so, what kind?)? Is there a pedestrian crossing and if so, what type?
- How? How did the collision occur? What are the contributory factors in the incidents? What are the pedestrian and related vehicle manoeuvres? Are the pedestrians crossing the road at the time?
- Who? Who are the pedestrians? What are their ages and gender? Are they from Liverpool or elsewhere? What is their background? (We link the postcodes of casualties and drivers from STATS19 to socio-demographic data – more on that below). Who are the drivers or riders who hit the pedestrians? What vehicle are they driving or riding? What are their ages and gender and are they local to Liverpool? What is their journey purpose? What is their background?
When attempting to answer the above questions, we look to make comparisons with other areas to see if what is happening in Liverpool is typical. If the risks facing adult pedestrians are the same across the country, then lessons could be learnt from elsewhere. If something unusual is happening in Liverpool, it could provide a new focus for action. In this project, adult pedestrian risk in Liverpool is being compared to a set of local authorities that have similar road networks and another set of local authorities which have a similar socio-demographic composition. This will allow us to understand if the focus should be on the road network or on the people using it.
Thinking about socio-demographics, understanding our target audiences is really important. MAST links postcodes of participants to other national statistics, such as rurality and the Index of Multiple Deprivation classification systems. It also links to Mosaic Public Sector, which is a classification system based on 450 different data sources where each postcode in the country belongs to one of 15 Groups and 66 Types. These Groups and Types are generalisations about the types of people who live in those communities and provides information on household composition, finances, work lives, education, health, crime and communication preferences.
All of this information is really important, understanding the trends in how the adult pedestrians were injured will guide whether the solutions may be engineering, educational or enforcement-based.
The detailed STATS19 analysis will guide the formation of those solutions, and the insight into who the participants are will help shape how those solutions are communicated to citizens. The Mosaic insight can guide the language and tone used, the partners used to deliver the message and what behaviour change levers might be most effective with them (remember, we are talking about the adult pedestrians and the drivers who hit them – there could be changes to their behaviour that both parties need to make).
The final part of this phase is to supplement the analysis. Are there other datasets which can explain the findings (such as understanding some of the social and economic challenges facing citizens or enforcement data related to driver behaviour)? Are there specific locations where risk is particularly high and if so, why?
In approximately two weeks time we will be in a position to share the headline findings from our data. But this will only tell us so much. To build up a true picture, So-Mo will be out on location looking to understand at first hand what how pedestrians and drivers are behaving in these locations and how the environment is influencing this.
I'd like to know how other areas are using data to shine a light onto the who, what and where of pedestrian casualties? Let us know by joining the conversation @So_mo_co #LpoolPedestrian