ESG Analysis of Replacing Euro 3 Vehicles with used Euro 6+ Models

PACKAGE CONTENTS This package contains: - the final thesis document in PDF and editable Word format;- the Python Jupyter notebooks used for data processing, analysis, and visualisation;- the R script used for the econometric analysis;- the public and distributable datasets used by the Python notebooks;- the figures generated during the analysis. The original vehicle-registration dataset used for the econometricmodels is not included because of its size and distribution conditions. Instructions for obtaining and importing this dataset are provided below. 2. PYTHON ANALYSIS The Jupyter notebooks included in the root directory are: - Fleet-EU-Accidents. ipynb- Social stats. ipynb- digitize-graphs. ipynb- hellasdirect. ipynb- national emissions reported. ipynb The corresponding input files are stored in the DATA directory. Before running a notebook, verify that the file paths used in its importcells point to the corresponding subdirectory inside DATA. If the packageis moved to a different location, local file paths may need to be updated. Recommended software environment: - Python 3. 11. 14- pandas 2. 3. 3- NumPy 2. 4. 4- Matplotlib 3. 10. 7- Seaborn 0. 13. 2- SciPy 1. 17. 1- Jupyter Notebook The notebooks should be executed from the first cell to the last cell. The generated figures are also available in the graphs directory. 3. ECONOMETRIC ANALYSIS IN R The econometric analysis is contained in: - regression models. R The analysis requires the European Environment Agency vehicle-level CO2registration dataset. The dataset can be obtained from the EEA CO2 Cars data portal: insert stable EEA dataset page or retrieval URL Download the vehicle-level CSV file covering the years used in the thesisand store it in a local directory. 4. LOCAL DUCKDB SETUP The DuckDB database file is not included in this package. It is createdlocally when the R script is executed. The user must select: a) a local path for the DuckDB database file; andb) the local path of the downloaded EEA CSV file. At the beginning of regression models. R, replace the original paths withpaths valid on the user's computer. Example: duckdbₚath <- "C: /reproduction/lake. duckdb" eeacsvₚath <- paste0 ("C: /reproduction/", "EUEMISSIONREGISTRATIONDATA₂010₂023. csv") con <- DBI: : dbConnect (duckdb: : duckdb (), dbdir = duckdbₚath) The script imports the CSV file directly into DuckDB by creating theeeaᵣaw table. The user does not need to create the tables manually. The CSV path used inside readcsvₐuto () must also be replaced with thepath stored in eeacsvₚath or with another valid local path. 5. REQUIRED R PACKAGES Install the required packages before executing the script: install. packages (c ("DBI", "duckdb", "dplyr", "ggplot2", "scales", "reshape2", "car", "lmtest", "sandwich") ) The econometric script should then be executed sequentially from top tobottom. The script: 1. creates or opens the local DuckDB database;2. imports the EEA CSV file;3. converts the original character fields into numerical variables;4. classifies the emissions-testing protocol;5. creates Euro-standard proxy groups;6. harmonises fuel categories;7. prepares the NEDC and WLTP regression samples;8. estimates the baseline linear models;9. estimates the log-linear robustness models;10. produces the regression diagnostics and visualisations. 6. DATA CONFIDENTIALITY The package does not disclose confidential individual-level customerinformation. Where proprietary or restricted datasets were used, only the analysiscode, aggregated results, or distributable supporting material areincluded. Access to the original proprietary records may be subject topermission from the corresponding data owner. 7. OUTPUTS Figures generated by the Python analysis are stored in the graphsdirectory. Figures produced through the R econometric analysis were exported fromRStudio and are presented in the thesis document. Users can regeneratethem by executing regression models. R after completing the local EEA andDuckDB setup described above. 8. REPRODUCIBILITY LIMITATIONS Exact numerical reproduction requires the same version and release of thesource datasets used in the thesis. Public databases may be revised afterdownload, and therefore later versions may produce small differences inthe results. The proprietary source datasets are not redistributed. Consequently, analyses relying on those records can be inspected through the suppliedscripts and outputs but may not be fully rerun without authorised accessto the original data.