README.md

---
license: cc-by-nc-4.0
language:
- tr
task_categories:
- text-classification
pretty_name: Mapping Climate Policy Discourse in Turkey - A Topic Modeling Analysis of Parliamentary and Social Media Debates (2021-2025)
tags:
- sentiment-analysis
- text-analysis
- climate-change
- legal
- social-media
- public-policy
size_categories:
- 10K<n<100K
configs:
- config_name: default
  data_files:
  - split: climatelaw_analysis
    path: all_platforms_masked.csv
  - split: waterlaw_forecast
    path: su_kanunu_masked.csv
dataset_info:
  features:
  - name: Name
    dtype: string
  - name: Profile ID
    dtype: string
  - name: Date
    dtype: string
  - name: Likes
    dtype: float64
  - name: Comment
    dtype: string
  - name: Platform
    dtype: string
  - name: Post ID
    dtype: string
  - name: Topic
    dtype: string
  - name: Original Post?
    dtype: bool
  - name: Sentiment
    dtype: string
  - name: Argument
    dtype: string
---

# Mapping Climate Policy Discourse in Turkey - A Topic Modeling Analysis of Parliamentary and Social Media Debates (2021-2025)

## Dataset Summary

This dataset contains Turkish-language social media posts related to public discourse on environmental legislation in Türkiye, specifically focusing on the **Climate Law (İklim Kanunu)** and the **Water Law (Su Kanunu)**.

The dataset covers the period **2021–2025** and includes content collected from **Facebook, Twitter (X), Ekşi Sözlük, and DonanımHaber Forum**. It is designed to support research on public opinion, sentiment analysis, argumentation patterns, and policy-related discourse in social media.

### Contextual Usage
* **Climate Law Posts:** Primarily used for retrospective analysis, as the law has arguably been enacted or is in advanced stages.
* **Water Law Posts:** Intended for forecasting public reactions to upcoming bill proposals.

All data was collected from publicly available content and processed for **non-commercial research purposes only**.

---

## Supported Tasks
- **Sentiment Analysis:** (Labels provided via `savasy/bert-base-turkish-sentiment-cased`)
- **Argument Classification:** (Sub-topics identified via LDA)
- **Policy Discourse Analysis**
- **Social Media Text Analysis (Turkish)**

---

## Dataset Structure

Each row in the dataset represents a single social media post or comment.

### Data Fields

| Column Name | Description |
|:---|:---|
| `Name` | Anonymized user name |
| `Profile ID` | User identifier |
| `Date` | Timestamp of the post |
| `Likes` | Number of likes or reactions (if available) |
| `Comment` | The actual text content of the post |
| `Platform` | Source platform (Facebook, Twitter, Ekşi Sözlük, DonanımHaber Forum) |
| `Post ID` | Unique identifier of the original post |
| `Topic` | Keyword used for scraping (e.g., `iklim_kanunu`, `su_kanunu`) |
| `Original Post?` | Boolean indicating if the entry is an original post or a reply/comment |
| `Sentiment` | Sentiment label (positive/negative) predicted using `savasy/bert-base-turkish-sentiment-cased` |
| `Argument` | Sub-topic label identified using **LDA topic modeling** (See table below) |

### Argument Labels (LDA Topics)

The `Argument` column represents sub-topics of discussion derived from Latent Dirichlet Allocation (LDA) topic modeling.

| Argument Label | Description | Top Words (Examples) |
|:---|:---|:---|
| **Energy Policy** | Fossil fuels, carbon, and energy policies | *iklim, fosil, yakıt, karbon, enerji* |
| **Climate Crisis** | Crisis related to fossil fuels & politics | *klim, fosil, yakıt, türkiye, krizinin* |
| **Green Activism** | Protests and global activism | *fosil, yeşile, küresel, protesto* |
| **Intl Conferences** | COP and international summits | *iklim, konferansı, taraflar, birleşmiş milletler* |
| **Ministry Actions** | Actions by the Ministry of Environment | *çevre, şehircilik, bakanlığı, murat (kurum)* |
| **Urban Policy** | Urban transformation and TOKİ | *devlet, kentsel, toki, dönüşüm* |
| **Bill Proposal** | Parliamentary bill proposals | *tbmm, genel, kurulu, teklifi* |
| **Green Legislation** | Green Deal and adoption processes | *yeşil, mutabakat, kanun, kabul* |
| **Approval Process** | Legislative approval/acceptance | *kabul, tbmm, meclis, edildi* |
| **Emissions Trading** | ETS, carbon tax, greenhouse gas | *emisyon, sera, ticaret, karbon* |
| **Politics & Forests** | Political decisions on forestry | *orman, geri, çekildi, erdoğan* |
| **Enactment** | Official Gazette publication | *resmi, gazete, yürürlüğe, girdi* |
| **Policy Debates** | General debates on climate policy | *iklim, karbon, hayir, küresel* |
| **Water Mgmt** | Water management and irrigation | *sulama, gölü, suyu, yönetimi, taşkın* |
| **Agri Risks** | Agricultural risks (drought/rain) | *yağmur, tarımsal, tehdit, hasat* |
| **Weather Alerts** | Meteorological warnings | *meteoroloji, alarmı, şiddetli, acil* |
| **Drought Crisis** | Regional drought impacts | *kuraklık, su, tehlikesi, krizi* |
| **Food Security** | Food supply and natural disasters | *gıda, doğal, afet, korumak* |
| **Desertification** | Desertification and soil loss | *çölleşme, mücadele, toprak, erozyon* |
| **Forest Threats** | Global warming threats to forests | *ormanlık, yangın, tehdit* |
| **Env Awareness** | General environmental awareness | *ısınma, çevre, koruma, bilinç* |
| **Temp Rise** | Rising temperatures/Heatwaves | *sıcak, ısınma, derece, artış* |
| **Climate-Drought** | Climate change causing drought | *iklim, değişikliği, kuraklık* |
| **Fire Containment** | Status of forest fire control | *kontrol, altına, alındı, söndürme* |
| **Firefighting** | Active firefighting operations | *yangın, müdahale, havadan, uçak* |
| **Evacuations** | Evacuations due to disasters | *tahliye, zarar, can, kaybı* |

---

## Data Collection and Processing

* **Time Span:** 2021–2025
* **Sources:** Publicly accessible posts from Facebook, Twitter (X), Ekşi Sözlük, and DonanımHaber Forum.
* **Methodology:**
    * Data was scraped using topic-specific keywords (`iklim_kanunu`, `su_kanunu`).
    * **Sentiment Analysis** was performed automatically using the Hugging Face model `savasy/bert-base-turkish-sentiment-cased`.
    * **Argument Sub-topics** were extracted using unsupervised LDA topic modeling.
    * **Anonymization:** Usernames and profile identifiers have been anonymized to protect user privacy.

---

## Ethical Considerations

This dataset contains user-generated content from social media platforms. Although the data was publicly available at the time of collection:
* **Privacy:** Personal identifiers have been anonymized.
* **Usage:** The dataset should **not** be used to identify, profile, or target individuals.
* **Scope:** No private messages or restricted content were included.
* **Bias:** The dataset may reflect platform-specific biases and does not necessarily represent the full population of Türkiye.
* **Language:** Social media text may contain informal expressions, sarcasm, or offensive language.

Researchers are encouraged to follow platform-specific terms of service and ethical research guidelines when using this dataset.

---

## Citation

If you use this dataset in your research, please cite it as follows:

```bibtex
@misc{turkish-env-social-media-2026,
  title = {Mapping Climate Policy Discourse in Turkey - A Topic Modeling Analysis of Parliamentary and Social Media Debates (2021-2025)},
  author = {Yıldırım, Cansu Mine and Buldu, Murat and Akdağ, Hüseyin Emir and Kahraman, Sena},
  year = {2026},
  publisher = {HuggingFace},
  howpublished = {\url{[https://huggingface.co/datasets/webomurga/CMPE59U_Social_Media](https://huggingface.co/datasets/webomurga/CMPE59U_Social_Media)}}
}