dmytrobuhai commited on
Commit
e14dbf4
·
verified ·
1 Parent(s): dd6b0e0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +85 -3
README.md CHANGED
@@ -1,3 +1,85 @@
1
- ---
2
- license: mit
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ task_categories:
4
+ - text-classification
5
+ - zero-shot-classification
6
+ - sentence-similarity
7
+ language:
8
+ - en
9
+ - uk
10
+ tags:
11
+ - climate
12
+ - quality
13
+ - pollution
14
+ - air
15
+ - wartime
16
+ - war
17
+ pretty_name: 'Ukrainian Air Quality: Daily Pollution in Wartime'
18
+ ---
19
+
20
+ This dataset presents daily measurements of air pollution across **more than 35 Ukrainian cities** during the years 2024–2025, collected from official open government sources. Each row represents a specific measurement of an air pollutant (e.g., sulfur dioxide, ammonia, soot) in a given city on a particular day.
21
+
22
+ To enrich the contextual dimension of this dataset, we included open-source incident data on civilian harm resulting from missile and drone strikes.
23
+
24
+ #🟢 Tasks Using Pollution Data (pollution_dataset.csv)
25
+ - 1️⃣ Time Series Forecasting: Air Pollution Prediction
26
+ - 2️⃣ Anomaly Detection in Air Quality (Detect unusual pollution spikes that deviate from normal patterns)
27
+ - 3️⃣ Clustering of Cities by Pollution Profiles
28
+
29
+ #🔴 Tasks Using Civilian Harm Data (ukr_harm.csv)
30
+ - 4️⃣ NLP: Attack Event Categorization (Classify attack descriptions into predefined categories)
31
+ - 5️⃣ Location-Based Risk Modeling (Assess the likelihood of future attacks in specific cities based on past trends)
32
+
33
+ #🟡 Tasks Using Both Datasets Together
34
+ - 6️⃣ Impact of War on Air Quality
35
+ - 7️⃣ Predicting Civilian Harm from Air Pollution Data
36
+ - 8️⃣ Spatio-Temporal Heatmaps of Pollution and Attacks
37
+
38
+ # 📂pollution_dataset.csv
39
+ Daily city-level air quality data from official Ukrainian sources
40
+ | Feature | Description |
41
+ | --- | --- |
42
+ | Date | Observation date (YYYY-MM-DD) |
43
+ | City | Ukrainian city name (in Ukrainian) |
44
+ | CoordinateNumber | Station coordinate ID |
45
+ | NameImpurity | Name of pollutant (translated to English) |
46
+ | Value | Daily measurement of pollutant |
47
+ | Civharm_id | Optional ID of conflict incident (if any) **by date** |
48
+
49
+
50
+ 📌 Source: [UKRAINIAN HYDROMETEOROLOGICAL CENTER OF THE STATE EMERGENCY SERVICE OF UKRAINE](https://data.gov.ua/dataset/63cd8827-e6f1-4e0f-8fa1-96f1a5872d9e)
51
+
52
+ 📎 License: Creative Commons Attribution 4.0
53
+
54
+ 📅 Coverage: Jan 2024 – Feb 2025
55
+
56
+ 🗂 Format: Compiled from 14 CSV/XLSX monthly reports
57
+
58
+ # 📂 ukr_harm.csv
59
+ Crowdsourced database of attacks impacting civilians during the war
60
+
61
+ | Feature | Description |
62
+ | --- | --- |
63
+ | id| Unique identifier for the incident |
64
+ | date | Date of the attack |
65
+ | location | City or region where the attack occurred |
66
+ | description| Short summary of the incident (may include civilian harm references) |
67
+ | type_of_area_affected | List of affected infrastructure types (e.g. Residential, Healthcare, etc.) |
68
+ | weapon_system | List of weapon types used (if available) |
69
+
70
+ 📌 Source: [Bellingcat's Ukraine Civilian Harm Map](https://ukraine.bellingcat.com/).
71
+
72
+ 🧾 Descriptions are based on open-source visual evidence (social media, geolocation, and video analysis)
73
+
74
+ ⚠️ Disclaimer: Content may include references to violence or casualties; viewer discretion advised.
75
+
76
+ ### 🔍 Data Cleaning & Standardization
77
+ Geolocation data was removed from ukr_harm.csv to protect privacy.
78
+ Non-essential metadata was excluded.
79
+ City names were standardized to match available locations in pollution_dataset.csv.
80
+
81
+ ## ⚠️ Ethical Notice
82
+ This dataset was compiled for public research and academic analysis.
83
+ The conflict layer may reference sensitive content, including descriptions of civilian casualties, destruction of infrastructure, and military-related terms.
84
+
85
+ ➡️ Use respectfully and responsibly.