Skip to content


Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Machine learning for research on climate change adaptation policy integration

Binder DOI DOI

This is the software repository accompanying the paper Machine learning for research on climate change adaptation policy integration: an explorative UK case study, by Robbert Biesbroek, Shashi Badloe, Ioannis N. Athanasiadis, published in Regional Environmental Change (2020).

This reposiroty contains all the software for reproducing the paper, and some sample documents. It can be executed online as a binder notebook, simply from your browser by clicking this button Binder.

Due to its size, the full dataset has been made available separately via Zenodo, here. The software comes with a subset of the full dataset, and can be demonstrated without any additional changes.


Do not change the names of any folders or files before finishing the pipeline, as the scripts look in folders with these specific names.

Step 1:

In the folder 'PDF_files' the following PDF documents are contained.
'Adaptation policy documents' - Training data for adaptation policies
'Mitigation policy documents' - Training data for mitigation policies
'Non-climate policy documents' - Training data for non-climate documents
'Mixed policy documents' - Testing data, any PDF document(s) you want to predict on.

Step 2:
In the folder 'Python Scripts' every script in the pipeline is contained.

Main pipeline - Extract raw text from PDf documents (parsed_files) - Filters, cleansand structurizes data into 'bags-of-words' (structured_files) - Builds database and inserts cleaned data (climate.db) - Builds vocabulary from training data (conversion_dictionary.txt) - Builds neural network and stores it (tensorflow/logdir) - Uses stored model to predict on new data. Results stored in database.

Optional scripts - Runs every script in the main pipeline in order. - Retrieves documents from website (PDF_files\Scraped documents) -- OUTDATED - Plots distribution of block lengths (Plots) - Plots histogram of test set blocks and their labels. WARNING: Only use few documents - Visualization for fraction of high confidence blocks - Launch tensorboard from stored model


No description, website, or topics provided.






No releases published


No packages published