X197 Construction of Data Set for Leakage calculations.

Tender description

Why the Work is Being Done Water and waste water treatment plants are extremely complex and expensive operations to run. We constantly monitor all aspects of these networks to ensure that our drinking water is top quality, plus our waste water plants run efficiently, adhering to stringent compliance regulations. This involves collecting vast amounts of data from Spatial, financial, telemetry and SCADA systems. Therefore, the data sets used to generate reports and business insights need to be up to date, be of excellent quality, plus be accurate. A lot of work has already been completed at NI Water to try to work with this data however, we need assistance from expert data analytics companies, to ensure that we make the most efficient and effective use of that data to learn how our networks are operating and how we can realise efficiencies. We have already used some of this data to realise some gains (financial) by finding areas of energy reduction in our water treatment plants but we are fully aware we can improve efficiencies by employing the latest and greatest algorithms and processes. We have many projects to complete on the water network but also on the waste water network. Problem to Be Solved NI Water needs to account for all water delivered to help determine if it is consumption or leakage. Each property is classified as Domestic, or non-Domestic however there may be data inaccuracies. Non-domestic property data is thought to be correct but the same degree of confidence does not exist for domestic properties. This means our leakage figures may be unnecessarily high and hence the need to determine we have the correct number of properties listed. Our objective is therefore to use as many data sets as possible to cross reference properties in NI to: 1.Identify anomalies in current property data , using all data sources available. 2.Produce creative methods to enable expeditious cross referencing of data from multiple sources . 3.Determine property counts and correct property classifications for all properties in NI. 4.Identify properties not on any system (using satellite imagery). 5.Determine count of properties updated (dashboard). 6.Communicate changes required (GIS/ Metering and billing). 7.Determine connection points for each domestic property that exist on GIS/ RAPID etc. 8.Propose how updates will be made to data sources. Success will require identification of all properties that could be updated, or provide justificatios to change how total leakage figure is calculated. Who Are the Users • Head of Water – Do we have all properties/ customer supply points listed and are they correctly classified in terms of domestic and non-domestic. • Head of Customer Services – need to ensure we have all customers updated to provide domestic and operational services. • Head of Business Intelligence - how can I see an accurate, up to date list of findings from this activity? Is it significant enough to change internal processes? Early Market Engagement There have been a lot of investigations in this area via internet searches and talking to data analytics companies. There is a real drive to combine as many data sets as possible together, to produce a definitive list we can use going forward. There have been a lot of creative ideas in how to find outliers which has prompted the urgency in preparing this piece of work. Work Already Done Our in-house GIS team have already developed a fantastic suite of applications to display layers of information in graphical format, utilising the latest in GIS technology and data sets (Pointer). We still however have a need to ensure that properties have the correct classification and we have an accurate listing of properties. This information will be made available as needed, provided the right data sharing agreements and contracts are in place. Existing Team Head of Analytics Head of Water Business Improvement Project Manager Business Analytics Project Manager Data Scientist Data Engineer Customer Insights Manager Current Phase Not started Work Location Northern Ireland Water Westland House Head Office 40 Old Westland Road BELFAST BT14 6TE Working Arrangments • Work as one team with the NIWater Analytics delivery team, whether on-site or off-site. • Provide regular updates on progress, issues, risks and delivery milestones. • Ensure that there is no duplication of work, enable shared learning and knowledge to NIWater • Ensure models are built in a manner that maximises ease of use/maintenance by NI Water once project is underway, be empowered to make suggestions for project updates as we learn more. • Flex resources as required, and work with other partners • Comply with NI Water policies/procedures. • Integrate with NI Water/ Microsoft tools already in place Security Clearance Basic Access NI check is required. On site working as required however current coronavirus guidelines should be adhered to at all times so remote meetings will always be preferred. Data sharing agreements will be required as appropriate, sensitive data obtained should be anonymised to adhere to strict GDPR regulations. Additional T&Cs NI Water intends to use the eTendersNI platform to evaluate this procurement. Shortlisted suppliers will be invited to submitted their responses on their prosposal, cultural fit and price via this platform. Scoring mechanism as follows will be used during both stages of the evaluation. 5 - Exceeded 3 - Met 1 - Partially met 0 – Not met or not evidenced NI Water do not envisage any additional terms and conditions Skills & Experience Demonstrate experience of successfully identifying domestic properties using Pointer (LPS) and other data sources to confirm property is completed (not under construction) and potential occupancy status. (10 points) Demonstrate experience using government data sets, or other approved data sets to build one key data set (e.g. NIE data set, Customs House data set). (10 points) Demonstrate experience of designing and implementing AI solutions to deliver operational efficiencies, service improvement and revenue (10 points) Demonstrate experience of delivering Agile projects within the past 3 years (10 points) Demonstrate experience building dynamic, drill down and static Reports and Dashboards for consumption on desktop, plus optimised for mobile devices. (10 points) Demonstrate experience of database concepts, design principles and database architectures. (10 points) Demonstrate experience of developing data-pipelines (10 points) Demonstrate experience of successfully developing reusable predictive analytical models to detect and predict fraud/error (10 points) Demonstrate experience of creating, maintaining and monitoring AI / ML solutions (10 points) Demonstrate experience of developing and deploying advanced analytics tools/ Azure cognitive services (diagnostic, predictive) using Python or R (10 points) Demonstrate experience building analytic models that blend GIS datasets and non GIS datasets (10 points) Demonstrate experience of using satellite imagery to cross reference against LPS. (10 points) Demonstrate experience using graph databases – interrogating relationships between nodes (10 points) Demonstrate experience of data matching processes and algorithms including fuzzy matching (10 points) Demonstrate experience of successfully integrating and managing multiple internal and external data sources (>20) on a single production-level platform (10 points) Demonstrate experience of working with address data in Northern Ireland to identify building classifications and methods on how to keep this updated automatically. (10 points) Demonstrate experience of successfully upskilling/transferring knowledge to the internal customer's team in relation to predictive analytics (10points) Demonstrate experience of successfully working collaboratively with a client as part of a joint team to detect errors/ fraud in data, essentially as an extension of a development team. Demonstrate experience of developing accurate project plans, clearing highlighting critical milestones and adhering to timelines and budget. (10 points) 28. Must be able to start work by end of May 2020, attend meetings in Westland House if necessary, although remote meetings will be accepted via Microsoft Teams. (10points) Nice to Haves Demonstrate experience of successfully developing predictive analytic models in the public sector (10points) Demonstrate experience with applications like Netbase, or similar. (10 points) Demonstrate experience of working successfully in several different industries/sectors (10points) Demonstrate experience working with APIs (10 points) Demonstrate evidence by example of using open source components to deliver a solution to detect and predict error/ fraud. (10 points) Demonstrate experience of successfully developing predictive analytics models that are similar to water industry where trends and predictive analytics are fundamental E.g.financial or utility sector are acceptable.(10points) No. of Suppliers to Evaluate 3 Proposal Criteria Identify a RAID (Risks, Assumptions, Issues, Dependencies) list that needs to be considered as part of the plan. Describe how you propose to ensure you have the right data sets to help us build a common data set that has a “A*” grading Description of how you have determined house occupancy which may skew the current assumptions. Description of how you will identify all stakeholders in these activities. Description of how you will use our current data and what insights could be gained in the early stages. Description of what kinds of dashboards you could implement to be able to display anomalies found. Describe of how you will build the models that will drive machine learning to help us become more proactive in keeping records up to date and maintain the application. Description of how we can migrate information to the cloud (if not already available) so we can support this ourselves going forward. Describe how you will provide knowledge transfer to up-skill the NIW team. Description of how you will document procedures, results as we complete projects. Description of how the models will be tested to prove that they are actually working and providing the right level of information. Proposed team structure you will require including experience (CVs) Cultural Fit Criteria How will you bring us to the next level in analytics? (10 points) What drives your organisation to deliver and, how will you ensure collaboration within the team? (10 points) How can you clearly demonstrate that you have the right solutions for us? (10 points) How can you ensure that you have dedicated resources for the entirety of this engagement and remain within project constraints? (10 points) How will you provide knowledge transfer to our in-house staff? (10 points) Payment Approach Capped time and materials Assessment Method Presentation Evaluation Weighting Technical competence 65% Cultural fit 15% Price 20% Questions from Suppliers No questions have been answered yet Budget range £70,000 - £90,000