CDP Analysis

Extracts CDP (Carbon Disclosure Project) questionnaire responses from corporate PDFs and city Excel files, normalizes them into JSON + Excel, builds analysis notebooks and reports, and serves an interactive dashboard on Streamlit.

Dashboard

Live dashboard (St. Petersburg): https://cdpanalysis-stpetersburg.streamlit.app/

Interactive exploration of St. Petersburg's CDP Cities responses — hazard/emissions data, persistent-blank causes, and question-theme breakdowns. Light mode by default.

• st_petersburg_dashboard.py — the Streamlit app
• st_petersburg_config.py — labels, colours, and the persistent-blank cause mapping (hand-edited only here)
• streamlit_dashboard_spec.md — original dashboard spec
• .streamlit/config.toml — forces base = "light" theme
• requirements.txt — pandas, openpyxl, streamlit, plotly

What's in the repo

Corporate CDP questionnaires (PDF → JSON → Excel)

Three companies share one pipeline cloned from apple_pdf_to_json.py. Two physical formats exist for each:

• Legacy years (C-prefixed codes, e.g. C6.1) — the C is stripped so id is numeric everywhere (6.1).
• New years (numeric codes, e.g. 7.6.1) — the CDP "Word version" corporate questionnaire export, which can also cover Water Security, Plastics and Biodiversity modules.

Company	Years	PDF source	Scripts	Outputs
Apple	2021–2025 (2021–23 legacy, 24–25 new)	apple_cdp_report/	apple_pdf_to_json.py, apple_json_to_excel.py	apple_cdp_report/Apple_CDP_*.json, Apple_Responses.xlsx
BMW Group	2021–2025 (2021–23 legacy, 24–25 new)	BMW_GROUP/	bmw_pdf_to_json.py, bmw_json_to_excel.py	BMW_GROUP/BMW_CDP_*.json, BMW_GROUP/BMW_Responses.xlsx
Unilever	2020–2024 (2020–23 legacy, 24 new)	unilever-cdp-report/	unilever_pdf_to_json.py, unilever_json_to_excel.py	unilever-cdp-report/Unilever_CDP_*.json, unilever-cdp-report/Unilever_Responses.xlsx

JSON schema (one file per year):

{
  "year": 2024,
  "source": "….pdf",
  "questions": [
    {
      "id": "7.6.1",
      "parent": "7.6",
      "section": "C7. Environmental performance - Climate Change",
      "question": "Gross global Scope 1 emissions (metric tons CO2e)",
      "row_label": "",
      "response": "55200"
    }
  ]
}

Excel workbook columns (long format, one row per question per year):

str, year, parent, section, row_label, question, response

Reports & notebooks

• Apple_CDP_Analysis.ipynb — Apple figures 1–6 and tables 1–2 from Apple_Responses.xlsx; full notebook of the underlying analysis.
• Apple_CDP_report.docx — Apple CDP analysis report.
• St_petersberg_cdp_analysis.ipynb + St. Petersberg CDP report.docx — St. Petersburg CDP Cities analysis.
• CDP_Mumbai_Analysis.ipynb — Mumbai CDP Cities extraction (see below).

Mumbai CDP Cities

Mumbai is a city, so its data lives in the CDP Cities dataset. CDP_Mumbai_Analysis.ipynb:

1. Installs dependencies (openpyxl, pandas).
2. Reads a CDP Cities Excel file (one sheet per question, wide format).
3. Keeps only rows where the disclosing organization is Mumbai.
4. Saves Mumbai's answers to a styled Excel workbook.

Configuration is in the notebook's Config cell:

Setting	Value	Meaning
INPUT_FILE	cdp_cities_data/2025_Full_Cities_Public_Data_Separated_by_Question.xlsx	CDP Cities input file
MUMBAI_DISC_NO	31178	Mumbai's cdp_disclosing_org_number
OUTPUT_FILE	Mumbai_Responses.xlsx	Output workbook
SKIP_SHEETS	{"Introduction", "Summary"}	Non-question sheets to ignore

Mumbai_Responses.xlsx — sheet "Mumbai Responses" with columns Question, Question Text, Sub-fields, Mumbai Response. Uses a dark blue header, alternating row shading, and text wrapping.

Setup

# 1. Create the environment and install dependencies
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt

# 2. Run a PDF -> JSON -> Excel pipeline
.venv/bin/python bmw_pdf_to_json.py
.venv/bin/python bmw_json_to_excel.py

# 3. Run the dashboard locally
.venv/bin/streamlit run st_petersburg_dashboard.py

How to run the notebooks

.venv/bin/jupyter lab CDP_Mumbai_Analysis.ipynb
.venv/bin/jupyter lab Apple_CDP_Analysis.ipynb

Run all cells in order.