To encourage data-driven decision making in public sector, University of Latvia and Microsoft Innovation Center developed a “Data challenge” platform and methodology. Aim of the initiative is to bring together teams of public sector officials and data analysts and in few weeks to create several AI solution prototypes for a specific public sector challenge giving the organisation an opportunity to explore innovative, data-driven solutions that they eventually can develop and implement in practice
Innovation Summary
Innovation Overview
Nowadays there is an immense pressure on public sector organisations to increase efficiency, decrease administrative burden on citizens and businesses, make wise data-based decisions while being cost effective and innovative all at the same time. From citizen perspective such demands make sense since government spends their tax money, thus they should work in the citizens’ best interests.
At the same time, public sector institutions have many legal and legislative restrictions hindering development and implementation of new, innovative yet sometimes less explored ideas, methods and programs. To invest money wisely in new ideas the institution has to be sure that the solution is worth it, yet they don’t have an opportunity to test it beforehand. This creates vicious circle and institutions get stuck.
Being an active part and one of the main facilitators of the innovation ecosystem in Latvia, the University of Latvia and Microsoft Innovation Center created a platform and a methodology called “Data challenge” to address the above-mentioned challenge. Data based problem solving and decision making is one of the challenges in public administration. It is therefore essential that both public sector professionals and data analysts exchange knowledge and gain experience in data analysis and data-driven decision-making. Every workshop is built around one well-defined public sector institutions problem. Teams of data analysts then try to create a solution various data analysis approaches, ranging from interactive data analysis to Machine Learning and Neural Networks
The target audience are the representatives from public sector and local business, data analysts and researchers from universities. During the seminar they are teaming up, working together to develop potential solutions for the defined task. At every event there is one public sector institution presenting a problem that they hope to be solved by creating a smart, data analytics-based solution. Teams comprised of data analysts, technology experts and enthusiasts then have a couple of weeks to explore available datasets and other relevant information and to create a solution prototype.
Then, the teams present their ideas and working prototypes at the next event where data analysis expert evaluates the model from technical point of view, and the organisation’s representatives give feedback on how useful and applicable the solution would be in the real setting.
Such platform and method give freedom and sparks enthusiasm to every involved party as it is set up as a challenge and a visionary method while still solving real life problems. Also, it is important that the holder of the platform is a neutral organisation building trust among everyone involved. The public sector organisations can see and assess the prototype before deciding to invest the funds in the solution development.
The “Data Challenge” seminars have already produced successful results. The Court Administration of Latvia participated in the “Data challenge” looking to improve efficiency of their processes. They provided data about the third category civil cases examined in the first instance in 2017. Using a variety of data analytics tools, a model (based on a neural network) was developed that could predict and determine the potential length of the court proceedings. The results provided by the data-driven test model were compared to the predictions of the judges about the same court proceeding cases, and it turned out that the judges were way more optimistic (and less precise than the test model) about how much time it takes for them to proceed the case. This solution now allows them to optimise their operations and organise their workflow much more precisely.
Similarly, Rural Support Service of the Republic of Latvia wanted to optimise the process of classifying images that local landowners and also their own employees take and use to evaluate the status of territory – whether the grass is mowed and removed or not. Based on the situation, Rural Support Service makes decision whether to grant support for the farmer. They receive huge number of photos and have to manually go through all those photos and mark each of them in one of the categories. It takes a lot of time and is not effective. Therefore, “Data challenge” teams used image recognition to offer a solution for this task. Now, Rural Support Service is considering the implementation of such solution that will save them time, money and will increase precision eliminating human error. Ultimately, landowners would be able to receive their state support faster improving their cashflow.
The “Data Challenge” platform and methodology can be applied to solving different problems – not only data-based ones. In future, this will be applied to different scenarios for public sector organisations on all levels.
Innovation Description
What Makes Your Project Innovative?
“Data challenge” is innovative because:
- A neutral platform allowing stakeholders involved in innovation development and implementation in public sector to meet, experiment, create prototypes without direct financial commitment.
- It enables both sides to focus on finding the best solution without any constraints. It has never been applied before and now it opens new opportunities to collaborate and develop innovative solutions.
- It is a rapid, cost and time efficient solution going from problem definition to working prototype within 4 to 6 weeks, and it requires only few man-hour investment from public sector representatives.
- Gives freedom of choice to everyone involved. The public sector organisation is not expected to choose any of the solutions if they do not see the added value or do not see it possible to adapt it at the given moment.
- Neither are data experts “tied” to the given challenge – if they do not see any particular challenge as worthwhile, they can choose not to solve it.
What is the current status of your innovation?
“Data challenge” currently could be associated with all of these stages – from identifying problems and opportunities to diffusing lessons, because the project is executed in an agile manner. The innovation is implemented and at the same time also evaluated after every event to make sure that it reaches the right goals and is not used as an end to itself but actually delivering value and new insights to everyone involved. As this model has already delivered good results, we constantly look for new organisations that could benefit from this workshop, but also, we explore new directions and new types of challenges the platform could be used to solve – expanding collaboration to other partners, organisations and types of problems.
Innovation Development
Collaborations & Partnerships
The University of Latvia and Microsoft Innovation Center – has developed the framework, involves the researchers, establishes and maintains relationships with all the parties.
Microsoft Latvia – provides technological expertise and technology professionals to support teams.
Organisation Riga AI, ML & bots meet-up - attracts technology partners, data scientists, leads initial discussions with public sector institutions and helps them to prepare the data and information for the challenge.
Users, Stakeholders & Beneficiaries
Customers of the Public sector institutions – improved experience when receiving services, tax money spent wisely.
The Institution – cost savings, reducing administrative burden, less paperwork, time effective, improved quality.
IT Industry partners and data analysts – business opportunities, interesting challenges, skills and expertise development.
Academics – opportunity to develop their competencies by using their knowledge to solve real-life problems.
Innovation Reflections
Results, Outcomes & Impacts
Since 2017:
- 6 public institutions have presented their challenge;
- 1 solution is under development and planned to be implemented;
- Several solutions are being discussed between public sector organisation and the technology partner;
- Participants from 44 institutions – 36 of those from public sector institutions, attended and actively participated in the events.
We conduct feedback interviews with public sector institutions presenting their challenge and every organisation admits that this methodology is unique in their experience and gives a chance to explore new, innovative and necessary solutions beyond their own capacity and existing opportunities. Organisations that have participated often recommend this event to others and also give suggestions which organisations could be other potential participants. Over time number of participants at the event steadily increases demonstrating the increasing popularity of the event – from 15 participants in the beginning to more than 60 currently.
Challenges and Failures
Lack of understanding of the added value is the key hindrance for the government institutions to participate and present their challenges. Explaining this takes time and patience on both sides, but communication here is key.
Fear that technologies will take their job – and the necessity to change this the mindset and find other opportunities is another challenge.
Bureaucracy and lack of support from top management in some cases - when everything is set up, the management slows down or even stops the process with one or another excuse (not the right time, etc.).
A lot of explanatory work and communication explaining the idea, the opportunities and the value of the “Data Challenge” platform is the key to success in such situations.
Lack of available and accessible data that are machine readable format to be used for analysis and to develop solutions.
Conditions for Success
The holder of the platform has to be proactive when searching for and talking to the public sector organisations – to spread the information about the challenge actively and through many relevant channels as it takes time and effort to inform institutions about new initiatives. Also, it is very important to keep close contact with all parties involved and to create communication circle with data analyst community on one hand and the public sector on the other. Strong leader who speaks both languages – that of analysts and public sector, here is the key.
On the side of public sector institution, several things have to be ensured. First, it is key that the necessary datasets are both available and accessible in machine-readable format. Sometimes to create more precise model data from other databases may be needed. Second, the organisation has to gain support from senior management. On top of that, strong sense of motivation and commitment is required.
Replication
The model has been replicated in the given framework for several times already (different institutions, different data-driven challenges). The innovation has not yet been implemented in other organisations nor has been yet replicated to address similar problems. However, we have recently started to work on adapting this platform create solutions for public sector challenges’ that are not data based. The platform and methodology are flexible and easily can be adjusted to other fields, problems and even industries if one should be interested to use this in private sector or to facilitate collaboration between only public sector organisations.
Lessons Learned
It takes time to create trust – public sector organisations are cautious with new, innovative but therefore unproven methods. Stability is highly valued, and they are not keen on jumping in the unknown. To show that your innovation is something one can rely on and that we will deliver expected results takes time from both sides. Everyone has to be open and honest with each other to make sure that we move in the right direction.
We also have to be agile and ready to adjust the model and platform to improve it – as a pilot project it has a lot of unpredicted flaws that have to be improved throughout the process. We have to be open about this when presenting the idea to the parties involved. Admitting that this is a process in progress and that some changes may happen over time but making sure that this does not cause discomfort for anyone involved and that you have contingency plans in place helps to build trust and also sets the right mindset for involved parties that they have to be ready for changes but it all will be taken care of and will be directed so that it eventually brings better results.
Commitment to go through the process form a – z and be ready to actively engage – although the method requires relatively little engagement and time investment form everyone involved, it is important that once the organisation has committed to participate, it has to go through the process till the end. It is important that the organisation has the higher level management support and no sudden exit from the project happens.
Strong leader who can hold the project “community” together – to organise public sector organisations, data analysts, technology experts and academics it takes one strong leader to find a way how to manage them. Creating dedicated communication platform and being available to anyone who may have any questions or support is crucial.
Find the right way to reach your audience – get the insiders to send out the information, then build trust.
Status:
- Identifying or Discovering Problems or Opportunities - learning where and how an innovative response is needed
- Generating Ideas or Designing Solutions - finding and filtering ideas to respond to the problem or opportunity
- Developing Proposals - turning ideas into business cases that can be assessed and acted on
- Implementation - making the innovation happen
- Evaluation - understanding whether the innovative initiative has delivered what was needed
- Diffusing Lessons - using what was learnt to inform other projects and understanding how the innovation can be applied in other ways
Date Published:
20 May 2019