electricsheepasia/asia-demographics-typhoon-mangkhut-2018-twitter-data

Typhoon Mangkhut 2018 Twitter Data

Publisher: Qatar Computing Research Institute · Source: HDX · License: cc-by · Updated: 2024-09-13

---

Abstract

This is a Twitter dataset collected during the typhoon Mangkhut 2018 in the Philippines. The data was collected, processed, and analyzed by the AIDR (http://aidr.qcri.org) platform using state of the art machine learning techniques. The data includes the reports of number of injured and dead people, infrastructure damage reports, missing or found people, urgent needs and donation offers for each hour. Due to Twitter TOS, we do not share full tweets content on HDX. Please contact us via HDX or on aidr.qcri@gmail.com to get tweet ids of the dataset along with a tool which can be used to rehydrate tweets from tweet ids.

Each row in this dataset represents time-series observations. Temporal coverage is indicated by the date column(s). Geographic scope: CHN, GUM, PHL.

Curated into ML-ready Parquet format by Electric Sheep Africa.

---

Dataset Characteristics

Domain	Humanitarian and development data
Unit of observation	Time-series observations
Rows (total)	244
Columns	16 (13 numeric, 2 categorical, 1 datetime)
Train split	195 rows
Test split	48 rows
Geographic scope	CHN, GUM, PHL
Publisher	Qatar Computing Research Institute
HDX last updated	2024-09-13

---

Variables

Geographic — sympathy_and_support_tweets (range 1.0–1640.0).

Temporal — date.

Demographic — affected_individual_tweets (range 0.0–581.0), infrastructure_and_utilities_damage_tweets (range 0.0–1069.0), number_of_images (range 0.0–133.0), mild_damage_images (range 0.0–38.0), severe_damage_images (range 0.0–24.0).

Outcome / Measurement — number_of_tweets (range 12.0–5518.0).

Identifier / Metadata — esa_source (HDX), esa_processed (2026-05-06).

Other — hour (range 0.0–23.0), injured_or_dead_people_tweets (range 0.0–164.0), missing_and_found_people_tweets (range 0.0–1.0), caution_and_advice_tweets (range 0.0–1104.0), donation_and_volunteering_tweets (range 0.0–113.0) and 1 others.

---

Quick Start

from datasets import load_dataset

ds    = load_dataset("electricsheepafrica/asia-demographics-typhoon-mangkhut-2018-twitter-data")
train = ds["train"].to_pandas()
test  = ds["test"].to_pandas()

print(train.shape)
train.head()

---

Schema

Column	Type	Null %	Range / Sample Values
date	datetime64[ns]	0.4%
hour	float64	0.4%	0.0 – 23.0 (mean 11.4691)
number_of_tweets	float64	0.4%	12.0 – 5518.0 (mean 584.2099)
affected_individual_tweets	float64	0.4%	0.0 – 581.0 (mean 50.6708)
infrastructure_and_utilities_damage_tweets	float64	0.4%	0.0 – 1069.0 (mean 69.3827)
injured_or_dead_people_tweets	float64	0.4%	0.0 – 164.0 (mean 7.9136)
missing_and_found_people_tweets	float64	0.4%	0.0 – 1.0 (mean 0.0206)
caution_and_advice_tweets	float64	0.4%	0.0 – 1104.0 (mean 81.679)
donation_and_volunteering_tweets	float64	0.4%	0.0 – 113.0 (mean 19.8971)
sympathy_and_support_tweets	float64	0.4%	1.0 – 1640.0 (mean 207.1728)
other_useful_reports	float64	0.4%	0.0 – 517.0 (mean 57.1317)
number_of_images	float64	0.4%	0.0 – 133.0 (mean 12.3416)
mild_damage_images	float64	0.4%	0.0 – 38.0 (mean 2.3827)
severe_damage_images	float64	0.4%	0.0 – 24.0 (mean 1.7119)
esa_source	object	0.0%	HDX
esa_processed	object	0.0%	2026-05-06

---

Numeric Summary

Column	Min	Max	Mean	Median
hour	0.0	23.0	11.4691	11.0
number_of_tweets	12.0	5518.0	584.2099	105.0
affected_individual_tweets	0.0	581.0	50.6708	3.0
infrastructure_and_utilities_damage_tweets	0.0	1069.0	69.3827	8.0
injured_or_dead_people_tweets	0.0	164.0	7.9136	1.0
missing_and_found_people_tweets	0.0	1.0	0.0206	0.0
caution_and_advice_tweets	0.0	1104.0	81.679	8.0
donation_and_volunteering_tweets	0.0	113.0	19.8971	10.0
sympathy_and_support_tweets	1.0	1640.0	207.1728	38.0
other_useful_reports	0.0	517.0	57.1317	6.0
number_of_images	0.0	133.0	12.3416	2.0
mild_damage_images	0.0	38.0	2.3827	0.0
severe_damage_images	0.0	24.0	1.7119	0.0

---

Curation

Raw data was downloaded from HDX via the CKAN API and converted to Parquet. Column names were lowercased and standardised to snake_case. Common missing-value markers (N/A, null, none, -, unknown, no data, #N/A) were unified to NaN. 14 column(s) were cast from string to numeric or datetime based on parse-success rate (>85% threshold). The dataset was split 80/20 into train and test partitions using a fixed random seed (42) and saved as Snappy-compressed Parquet.

---

Limitations

• Data originates from Qatar Computing Research Institute and has not been independently validated by ESA.
• Automated cleaning cannot correct for misreported values, definitional inconsistencies, or sampling bias in the original collection.
• This dataset spans 3 countries; geographic and methodological inconsistencies across national boundaries may affect cross-country comparability.
• Refer to the original HDX dataset page for the publisher's own methodology notes and caveats.

---

Citation

@dataset{hdx_asia_demographics_typhoon_mangkhut_2018_twitter_data,
  title     = {Typhoon Mangkhut 2018 Twitter Data},
  author    = {Qatar Computing Research Institute},
  year      = {2024},
  url       = {https://data.humdata.org/dataset/typhoon-mangkhut-2018-twitter-data},
  note      = {Repackaged for machine learning by Electric Sheep Africa (https://huggingface.co/electricsheepafrica)}
}

---

Electric Sheep Africa — Africa's ML dataset infrastructure. Lagos, Nigeria.