Skip to content

Urban Narrative - AI Augmented Approach to Identify Shared Ideas from Large Format Public Consultation

License

Notifications You must be signed in to change notification settings

samminweng/urban_narratives

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

25 Commits
 
 
 
 
 
 
 
 

Repository files navigation

AI Augmented Approach to Identify Shared Ideas from Large Format Public Consultation (Urban Narratives)

This repository provides a prototype of a demo system (backend and frontend) to construct a demo website that visualise and explore large amount of public comments collected from a public consultation exercise called “Share an Idea” undertaken in Christchurch New Zealand after the 2011 earthquake.

Our implementation comprises backend and frontend. The backend is implemented in a mix of Python and Java to extract linguistic features from the texts with Stanford NLPCore toolkit (v.4.2.0), and collect and analyze the linguistic features. Stanford NLPCore toolkit is a Java NLP framework that provides commonly used natural language processing functionalities. In our case study, we make use of its sentence splitting, word tokenizer, part-of-speech tagger, lemmatization, dependency parser, and open information extractor (see Stanford NLPCore toolkit). The frontend is implemented in Javascript to provide a visualisation of analyzed results from the backend by using a set of open source libraries (Bootstrap, JQuery, Google chart, D3 and Plotly).

In terms of data workflow, the backend takes raw texts collected from the dataset as inputs and extracts linguistic features from the texts with Stanford NLPCore toolkit and then identifies core words and aggregates texts, linguistic features as output. The output of the backend is stored as a file in JSON format for data exchange. The frontend loads the JSON file from the backend, populates the dataset and feeds into the graphical library (D3, Google chart, Plotly, etc) to produce the interactive visualisation on the web page.

Required Software

  • Chrome
  • Python (3.6)
  • Java (1.11)
  • Stanford NLPCore Toolkit (4.2.0)
  • Python packages: pandas, stanza, tqdm

Build Instruction

Please refer to README in backend and frontend folders respectively.

Publication

If you are interested in our project and want to know about it, see our paper:

Weng M-H, Wu S, Dyer M. AI Augmented Approach to Identify Shared Ideas from Large Format Public Consultation. Sustainability. 2021; 13(16):9310. https://doi.org/10.3390/su13169310

About

Urban Narrative - AI Augmented Approach to Identify Shared Ideas from Large Format Public Consultation

Resources

License

Stars

Watchers

Forks

Packages

No packages published