The result of that effort is known as the Plastic Recovery Insight and Steering Model (PRISM). Co-created with IBM Consulting, PRISM fully realizes the Alliance’s vision of a secure place for stakeholders to convene, collaborate and innovate. The platform’s core function is to give stakeholders the data and tools they need to analyze and prioritize project opportunities anywhere in the world. To that end, it presents information to users visually, in the form of heat maps that display critical data, such as plastic leakage and waste processing capacity, at a granular geographic level.
Visual mapping is valuable because it provides decision-makers with a framework for assessing the high-level parameters of the situation in a particular region, country or city. But when it comes to making the business case for taking action, digging down into granular, high-quality data is a must. That’s why Sabine Strnad, an advisor to the Alliance who was leading the PRISM project, sees the built-in governance processes developed by IBM Consulting as the most critical aspect of the PRISM solution.
“When a report comes out, the fact that there’s no standardized way of reporting on plastic waste means there’s a lot of variation in the small details like what kind of plastic is included, what was the research used and how many households were surveyed,” says Sabine. “Failure to systematically take those differences into account undermines data credibility.”
Under the PRISM process, the proposed governance structure consists of a Governance Council, whose role is to drive strategic direction for PRISM, and to define what kind of data goes into PRISM and what new features and capabilities get built into it. Within the council are different working groups focused on data quality standards, policy and methodology, and technology.
In addition, before a data set is ingested into PRISM, a Review Group of subject matter experts examines it to understand the baselines and assign data quality scores to the data. For a community that’s committed to UN Sustainable Development goals and rightly vigilant against the infiltration of “greenwashed” data, this rigorous process helps build the trust that’s so essential.
What about when there’s a literal data gap, when a data element needed to make an investment decision doesn’t exist? It’s a common issue, with the estimated missing share of data points ranging from 60% to more than 90%. PRISM’s answer is to apply machine learning algorithms — along with augmented data sources related to economics and demographics — to fill in the gaps.
PRISM uses machine learning algorithms in IBM Watson® Studio to create archetypes of cities and regions, which are then used to estimate plastic leakage for a given place. Some of the important factors that go into this algorithm are proximity to coast, tourist population, run-off coefficient, GDP per capita, population density and policies around plastic waste management.
Using neural network-based algorithms, the Alliance has been able to model plastic leakage information for many cities in developing countries like India and Indonesia. A cloud-native solution, PRISM runs on IBM Cloud®, with the front-end portal running on IBM Cloud Foundry and Kubernetes clusters. The fact that PRISM runs on IBM Cloud means it can scale as the volume of data and users grows.