Given the dataset submitted by individuals on their ideas of how companies can implement the circular economy in their businesses, the team developed an algorithmic tool to validate these ideas against certain keyword metrics.
main.py: keyword metrics can be modified here and running will write into scores.txt the scores of each entry (not including spam and poor solutions) for each keyword metric and topsentences.txt the key sentences in each solution on each keyword metric graph.py will output the scores.txt into a graph to see the distribution of scores to help visualize what high percentile scores are vs low percentile scores ranks.py will output the n-best scores based on the scores in each category CSVreader.py shows each entry of problem and solutions one by one on key input
Sustainable Scan evaluates each solution to a sustainability problem based on the following rationale metrics: maturity stage, market potential, feasibility, scalability, technological innovation or adherence to circular economy principles. This code uses a branch of AI (NLP) to analyze the solutions with the NLTK tool and Word2Vec Algorithm.
Sustainable Scan ranks these solutions through three main steps:
- Processing - orginizes data by the given index, problem, and solution
- Filtering - removes solutions that are considered "bad" or spam.
- Scoring - Ranks the correlation of each word to the keywords [Sustainable, Feasible, Innovative, Novelty] The higher the accumulative score, the higher the correlation and more strong the solution is.
- Determines which index (solution) is the strongest based on its 4 categories (Sustainability, Feasiblilty, Innovation, Novelty).
- Showcases the distribution of the solution scores based on its categories with a graphical representation