List of Questions

How could a concept for a data shop look like?

Background information will be provided to academics signed up to this question prior to the event.

How can we harness the power of data with regards to fashion trends and production?

Background information will be provided to academics signed up to this question prior to the event.

How can we harness the power of text and voice recognition?

Background information will be provided to academics signed up to this question prior to the event.

How Can Data Science Link Emission Data to Air Quality Data from Sensors in Cities?

There is uncertainty in determining how much of the Air Pollution measured at Air Pollution Sensors can be attributed to specific traffic at specific roads or specific vehicle emissions. It has gained extra uncertainty and attention from inaccurate vehicle emission datasets through the ‘vw scandal’.

Despite this inaccuracy linking Air Pollution from Emissions from Vehicle on real world driving i.e regular road use to measurements at Air Pollution Sensors is a complex and uncertain task. It is a highly under studied question by Data Science methods with the possibilities to bring social, environmental and health benefits through insights gained.

Could Data Science be used to link and interpolate available datasets or even predict Air Quality producing other insightful datasets? There is large benefit in reducing the uncertainty in available datasets for knowing specific health exposure and to accelerate the modernisation of vehicles fleets. There are many older vehicles still in use both in public and private fleets of which there is a significant expense to replace which without government grants wouldn’t be replaced quickly enough to meet government own regulation limits.

It is clearly stated by government its plan is to regulate through clean air zones not have rapid replacing of fleets. The management of these clean air zone to insure success is very important and a large opportunity for reducing uncertainty in available datasets.

The pushing of larger emissions to other zone outside these means older and larger emission vehicles shall congest other zones. Most likely these zones shall have an increase in Air Pollution.

Being able to reduce uncertainty in Air Pollution Sensors and links to Emission from these vehicles would give strong evidence to reduce these vehicles from the fleet instead of just removing them from a few square km of our cities.


While running a Company, AirNode, which visualises Air Quality Data in Cities and produces Air Quality Modelling, I am very aware there are varied method to reduce the complexity of this problem. These include, though not limited, methods from validating datasets, increasing sensor usage, traffic modelling to Air Quality Modelling.

Although Data Science is the optimum method to reduce uncertainty of existing datasets, learn from previous datasets, test inferences and possibly forecast success of modification to cities.

There are existing although not very similar research projects some of are:

http://www.magic-air.uk/

https://www.kcl.ac.uk/lsm/research/divisions/aes/research/ERG/modelling/Air-quality-health-research.aspx

https://www.birmingham.ac.uk/news/latest/2018/01/birmingham-to-host-air-quality-research-supersite.aspx

https://www.gla.ac.uk/research/az/airpollution/projectoverview/

https://www.surrey.ac.uk/global-centre-clean-air-research

https://www.turing.ac.uk/research_projects/clean-air-london/

http://www.cam.ac.uk/research/features/bad-air-day-low-cost-pollution-detectors-to-tackle-air-quality

https://www.iscapeproject.eu/

https://www.pla.co.uk/Air-Quality-Strategy and

http://www.claircity.eu/ Data Science Dataset s Emission Analytics - Vehicle Emissions

http://equaindex.com/ Road side Sensor datasets

https://ricardo.com/news-and-media/press-releases/ricardo-exhibits-real-world-driving-emissions-test UK Government Air Quality Datasets

https://uk-air.defra.gov.uk/

How could we implement/set/create a standard data format across company without a huge manual effort?

As a global company split across numerous functions, data can appear in disparate formats. Data is stored in varying databases depending on function and/or region. Data ranges from the sourcing of raw materials to the production of finished goods, all the way to point of sale. How can we either implement a set structure, an ethos of uniformity or a best practice across the company without a large upheaval of current practices? How should we trace our path forward to an integrated data model, allowing the analysis of data across the company to gain greater and more thought provoking insights?

How can Data be verified and aggregated in simple, effective and automated process across a global company?

With current and historical data what sort of system/process can be envisaged to validate and align data to a set of predefined constraints? As a company we collect vast amounts of data that if not properly clarified initially will have to be restated after the fact. How can we remove or limit the human element to the restating of past data? Is there an easily introduced method to trawl through data and alert any discrepancies/outliers for further examination?

How can we promote and drive data ownership within our company?

With so many sources of data it can be difficult to ensure traceability across the entire company. How do we encourage the spirit of ownership among data collectors within the company? We wish to foster a spirit of pride and deep knowledge for each data owner, such the data issues are solved at the root cause. Have there be examples of shifting towards this approach to data collection?

How do we ensure Personally Identifiable Information (PII) remains secure and confidential within our company?

With cyber-attacks targeting personal information becoming ever frequent how do we safe guard our employees from this threat? How do we ensure sharing of PI within the company is safe from intrusion? This becomes an increasing concern as the line between work and home is blurred such that devices are often used for personnel and professional use. Without a heavy handed approach can we reassure employees that there information is held in a secure fashion.

How can we better use data to demonstrate the social outcomes of infrastructure investment?

Commonly we invest in infrastructure to facilitate a certain desired outcome. This might be faster journey times, increased energy outputs etc etc. However; for many infrastructure and development projects, both within cities and outside them, these projects also have the ability to deliver a wider range of social outcomes. Social inclusion, access to employment, a sense of belonging to a community etc.

The question is therefore about how we can better use readily available data (or bespoke data in certain circumstances) to more fully understand the extent to which recent projects have done this or future projects could. Also, how can we better use data to monitor these outcomes over time.

How will digital data alter our travel behaviour?

None.

Open Data - How do we encourage greater use of open data and how do we consume more of data that is available elsewhere

We have recently launched our CKAN Open Data platform for the NHS in Scotland (https://www.opendata.nhs.scot) and keen to understand how best to maximise use, and to understand how to use and consume Open Data from other sources

What are the industry standards or recommended approaches to data management and the data lifecycle that are suitable for large public sector organisation(s) who have a responsibility to source, manage, analyse and share data and information to a wide range of users? Are there any strategic tools available for this?

Very interested in understanding industry best practice for strategically planning for data and information across all aspects of accessing, managing and analysing data.

Very interested to have the support from researchers or academics about how we should organise and structure our data, and how we could visualise this. The main driver for this is the need to have a new vision and high level direction for our organisation which is responsible for all official statistics for the NHS in Scotland.

How do we organise clinical images and other data in a system that allows traceability / reproducibility of data flow through; data cleaning steps, quality assurance steps, and can interface securely to external parties who provide annotation data?

The data system must be operable by staff with no programming experience and to algorithm developers who have complex data categorisation needs. Systems must be compliant according to FDA guidance for industry such as “Computerized Systems Used in Clinical Investigations”.

What methods, statistical and organisational, can we use to assess the reliability of crowd sourced medical image annotations and labellings?

Annotations are of various types such as point based, region of interest, single or unspecified number per image. Reliability assessment may help us to prove the suitability of the non-medical workforce for the given task or to find and reject outlying performance. The assessment could be either internally or relative to medical expert annotations or labelling.

How to help users find which neighbourhood is best for them to live in

I would like to create a map visualisation with traffic light system of what different neighbourhoods are like for the UK and major expat cities (Dubai, Singapore, New York, Amsterdam, Sydney) using open source data and API's combining crime stats/ average house prices/average rent prices/school ratings/commuting distances. The user should be able to use sliders to determine which factors are most important to them to show which then changes the colours of the areas showing which are best for them.

What opportunities are there for academics and commerce to work together more effectively on environmental data?

Solutions to major environmental problems are likely to require the contribution of data scientists outwith the current scientific community, and with utilisation of commercial entities capabilities. What considerations need to be included in new structures / working practices that will be required?

How to best create a shared distributed blockchain where entire professional (education, project, work, testing) histories live and are validated?

Currently, reference checks are done following the entire interview process and just prior to onboarding. When these fail, this is a waste of time and money for companies (and job seekers). The ability to store education, projects, work history, publication history, and anything else that would be necessary to instantly identify an individual's track record, would vastly increase the ability to identify and instantly verify employment qualifications for every individual on the chain. How to do this on a global scale, from both a technical and theoretical standpoint, would improve the ability for individuals all around the world to have mobility for their careers and for companies to benefit from access to the best people on the planet to work for their companies.

How can we work with academia to improve the statistical models we have developed for property valuation?

Symmetry Analytics was created to use the power of advanced analytics to help consumers make better decisions, levelling the informational playing field between them and vendors. As an initial venture, we have developed a highly-predictive prototype model of residential property valuations, with a view to giving buyers and sellers of houses better, unbiased information. We seek to improve the model using techniques that might include spatial econometrics and image recognition. We are interested in collaborating with academia to do this better.

How do we build a holistic data management strategy that can be implemented on a practical basis?

This is a challenge for not only Edinburgh but the public sector, and offers an opportunity for discussion thinking on a national basis.

The City of Edinburgh Council is currently developing its thinking on models for data driven innovation projects that will be delivered through the Edinburgh and South East Region City Deal. As part of this work these questions have been developed to help build a better understanding of how to achieve this.

How do we create safe places to play with data?

In order to evolve understanding and value generation of our data we want a safe environment to explore the data.

The City of Edinburgh Council is currently developing its thinking on models for data driven innovation projects that will be delivered through the Edinburgh and South East Region City Deal. As part of this work these questions have been developed to help build a better understanding of how to achieve this.

How do we change data thinking in a public sector environment?

As the Council is developing innovation approaches to improve and re-design our services we need to identify ways to use our data creatively to support and influence this work, and build new ways of thinking to support it.

The City of Edinburgh Council is currently developing its thinking on models for data driven innovation projects that will be delivered through the Edinburgh and South East Region City Deal. As part of this work these questions have been developed to help build a better understanding of how to achieve this.

How can we develop tools so that managers can answer their own questions without needing experts?

In order to derive more value from our data and inform both policy and operational decisions we need to make it easier for officers to work with data.

The City of Edinburgh Council is currently developing its thinking on models for data driven innovation projects that will be delivered through the Edinburgh and South East Region City Deal. As part of this work these questions have been developed to help build a better understanding of how to achieve this.

How can we share data consistently across organisations and use cross sectoral data to inform community planning

To inform, influence and support community planning we need to be able to share data in a consistent manner across organisations. How can we create an environment and tools to do this?

The City of Edinburgh Council is currently developing its thinking on models for data driven innovation projects that will be delivered through the Edinburgh and South East Region City Deal. As part of this work these questions have been developed to help build a better understanding of how to achieve this.

How could we develop, as an open source project, an online voting platform based on blockchain that could be customised for many uses?

wallet.services is a 10-person start up focused on providing a platform to support the development of blockchain applications.
We have also been commissioned by Scottish Government to research and write a report on potential use cases for blockchain within the Scottish Public Sector.