# Exploring D3 Implementation Challenges on Stack Overflow This wiki page explains the contents of this OSF project repository. Please see the other pages for details on how to run the code, and how to apply it to Reddit data. Please see our [technical report on ArXiv](https://arxiv.org/abs/2108.02299) for more details regarding this research project. ## Manifest **Data** * `codebook.pdf` contains the codebook applied to our sampled Stackoverflow dataset, including definitions and examples of how Stackoverflow posts were qualitatively coded. * The original coded data is available in the `coded_data.xlsx` file. All dataset row numbers referenced in the paper are with respect to this file. * `other-apis-tools-analysis.xlsx` contains part of our analysis of different APIs and tools observed alongside D3. * For programmatic analysis, the coded data is translated to CSV files in the `code/codes` folder, stored as separate CSV files across the three coders (datsets D and E were coded by coder A, sorry for any confusion!). * All labeled D3 gallery images are in the galleryImages folder (located in `code/galleryImages.zip`). * `code/d3ExamplesUrls.json` contains the extracted urls for the observed examples in the D3.js gallery. These were used to search the Stackoverflow posts for known D3 example code. * Useful collections of mappings between documents (Stackoverflow posts) and codes are provided in `code/results_codes_map.json` and `code/results_docs_map.json`. These are the inputs used to execute our quantitative analyses. * Some intermediate files for mapping old codes to better formatted codes are located in `code/vis_codes_map.json` and `code/interaction_codes_map.json`. These intermediate codes should already be fixed in our coded data. They have not yet been phased out of the codebase, but are essentially redundant information. * All Stackoverflow html pages and contents of reddit posts are in the `code/pages` folder (located in `code/pages.zip`). A separate folder is also available for the reddit posts only for ease of use (`code/reddit.zip`, see notes on reddit data below). * `code/all_stackoverflow_urls.csv` contains the full list of Stackoverflow URLs. Note that some links may no longer be active. * `code/root_frequencies.json` and `code/call_frequencies.json` contain the data used to analyze the frequency of D3 operator calls in the full Stackoverflow dataset. **Code** * `code/analyzePostsHelper.py` contains core functions for our quantitative analysis of the full Stackoverflow and Reddit posts. * `code/analyzeCodes.py` performs our quantitative analysis of the qualitatively coded Stackoverflow posts. * `code/analyzeStackoverflowPosts.py` runs our quantitative analysis experiments for the full Stackoverflow dataset. * `code/analyzeRedditPosts.py` runs our quantitative analysis experiments for the Reddit dataset (see notes on reddit data below). * `code/extractAllCodes.py` extracts the qualitative codes, quotes, and links associated with each coded document from Stackoverflow (i.e., produces `code/results_codes_map.json` and `code/results_docs_map.json`).