Research datasets / Issue 2026-10

PEMRA complaints: provisional text labels and classification examples

Name: PEMRA complaints: provisional text labels and classification examples
Creator: Adaad

Complaint labels, classifier comparisons, co-occurrence, campaign totals and denominator exercises from Ticking more than one box.

Read the analysis →

Source and definitions

PEMRA Complaint Call Centre monthly reports; Adaad extraction, clustering, codebook and model outputs.

Complaint and distinct-text counts, shares, model scores, precision, recall and F1. The codebook and method note define each label.

pemra_complaint_labels.csv

Download CSV · 3477 rows

pemra_complaint_labels.csv — first 50 rows; download the CSV for the full table
uid	complaint_id	year	month	channel	lang	cluster	cluster_size	is_campaign	pemra_labelled	llm_target	llm_other	llm_abstain	llm_confidence	pemra_g_religion	pemra_g_hate	pemra_g_morality	pemra_g_misinfo	dict_g_religion	dict_g_hate	dict_g_morality	dict_g_misinfo	model_g_religion	model_g_hate	model_g_morality	llm_g_religion	llm_g_morality
2020-01-111273	111273	2020	1	A Plus Center	english	0	1	False	0	show	0	0	high	0	0	0	0	0	0	1	0	0	0	1	0	1
2020-01-111335	111335	2020	1	Abbtak	english	1	1	False	1	news	0	0	high	1	1	1	0	0	1	0	0	0	0	0	1	0
2020-01-111337	111337	2020	1	Abbtak	english	2	4	False	1	unclear	1	1	low	1	0	1	0	0	0	0	0	0	0	0	0	0
2020-01-111339	111339	2020	1	Abbtak	english	2	4	False	1	unclear	1	1	low	1	1	1	1	0	0	0	0	0	0	0	0	0
2020-01-111340	111340	2020	1	Abbtak	english	3	1	False	1	channel	1	0	low	1	1	1	0	0	0	0	0	0	0	0	0	0
2020-01-111341	111341	2020	1	Abbtak	english	4	1	False	1	news	0	0	low	1	1	0	0	0	0	1	1	0	0	1	0	1
2020-01-111343	111343	2020	1	Abbtak	english	2	4	False	1	unclear	1	1	low	1	0	1	0	0	0	0	0	0	0	0	0	0
2020-01-111344	111344	2020	1	Abbtak	english	5	5	True	1	news	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111345	111345	2020	1	Abbtak	english	5	5	True	0	news	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111346	111346	2020	1	Abbtak	english	6	1	False	0	news	0	0	medium	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111347	111347	2020	1	Abbtak	english	5	5	True	0	news	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111348	111348	2020	1	Abbtak	english	5	5	True	0	news	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111349	111349	2020	1	Abbtak	english	5	5	True	0	news	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111350	111350	2020	1	Abbtak	english	7	1	False	0	unclear	0	0	low	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111351	111351	2020	1	Abbtak	english	8	1	False	0	unclear	0	0	medium	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111352	111352	2020	1	Abbtak	english	9	2	False	1	unclear	0	0	medium	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111353	111353	2020	1	Abbtak	english	10	1	False	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111354	111354	2020	1	Abbtak	english	11	1	False	0	news	0	0	high	0	0	0	0	1	0	0	0	1	1	0	1	0
2020-01-111355	111355	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-11135	11135	2020	1	Abbtak	english	13	1	False	0	unclear	0	0	medium	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111357	111357	2020	1	Abbtak	english	9	2	False	1	unclear	0	0	medium	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111358	111358	2020	1	Abbtak	english	14	1	False	0	show	1	0	low	0	0	0	0	0	0	0	0	0	0	1	0	0
2020-01-111359	111359	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111360	111360	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111361	111361	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111362	111362	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111363	111363	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111364	111364	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111365	111365	2020	1	Abbtak	english	15	1	False	1	unclear	1	1	low	1	1	1	1	0	0	0	0	0	0	0	0	0
2020-01-111366	111366	2020	1	Abbtak	english	2	4	False	1	unclear	1	1	low	1	1	1	0	0	0	0	0	0	0	0	0	0
2020-01-111367	111367	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111368	111368	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111369	111369	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111370	111370	2020	1	Abbtak Center	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111371	111371	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111372	111372	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111373	111373	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111374	111374	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111375	111375	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111376	111376	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111377	111377	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111378	111378	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111379	111379	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111380	111380	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111381	111381	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111383	111383	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111384	111384	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111385	111385	2020	1	Abbtak	english	12	39	True	1	unclear	0	0	high	1	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111386	111386	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0
2020-01-111387	111387	2020	1	Abbtak	english	12	39	True	0	unclear	0	0	high	0	0	0	0	1	0	0	0	1	0	0	1	0

pemra_classifier_metrics.csv

Download CSV · 18 rows

pemra_classifier_metrics.csv — data as published
classifier	ground	label	n_pos	n	precision	recall	f1	tpr	fpr	tp	fp	fn	tn
dictionary	g_religion	Religion and sect	65	169	0.648	0.708	0.676	0.708	0.24	46	25	19	79
dictionary	g_hate	Hate and violence	61	169	0.556	0.492	0.522	0.492	0.222	30	24	31	84
dictionary	g_morality	Vulgarity and morality	62	169	0.717	0.613	0.661	0.613	0.14	38	15	24	92
dictionary	g_gender	Women and gender	16	169	0.484	0.938	0.638	0.938	0.105	15	16	1	137
dictionary	g_misinfo	False or defamatory	8	169	0.0	0.0	0.0	0.0	0.093	0	15	8	146
dictionary	g_state	State and security	4	169	0.167	0.25	0.2	0.25	0.03	1	5	3	160
model	g_religion	Religion and sect	65	169	0.54	0.723	0.618	0.723	0.385	47	40	18	64
model	g_hate	Hate and violence	61	169	0.828	0.393	0.533	0.393	0.046	24	5	37	103
model	g_morality	Vulgarity and morality	62	169	0.8	0.645	0.714	0.645	0.093	40	10	22	97
model	g_gender	Women and gender	16	169	0.64	1.0	0.78	1.0	0.059	16	9	0	144
model	g_misinfo	False or defamatory	8	169	0.0	0.0	0.0	0.0	0.012	0	2	8	159
model	g_state	State and security	4	169	0.0	0.0	0.0	0.0	0.018	0	3	4	162
llm	g_religion	Religion and sect	65	169	0.591	0.846	0.696	0.846	0.365	55	38	10	66
llm	g_hate	Hate and violence	61	169	0.829	0.475	0.604	0.475	0.056	29	6	32	102
llm	g_morality	Vulgarity and morality	62	169	0.796	0.694	0.741	0.694	0.103	43	11	19	96
llm	g_gender	Women and gender	16	169	0.577	0.938	0.714	0.938	0.072	15	11	1	142
llm	g_misinfo	False or defamatory	8	169	0.0	0.0	0.0	0.0	0.019	0	3	8	158
llm	g_state	State and security	4	169	0.0	0.0	0.0	0.0	0.018	0	3	4	162

pemra_threshold_curves.csv

Download CSV · 114 rows

pemra_threshold_curves.csv — first 50 rows; download the CSV for the full table
ground	threshold	precision	recall	fpr	raw_share	adjusted_share	true_share_gold
g_religion	0.05	0.4038	0.9692	0.8942	0.8208	0.0	0.3846
g_religion	0.1	0.4234	0.8923	0.7596	0.7245	0.0	0.3846
g_religion	0.15	0.4419	0.8769	0.6923	0.6537	0.0	0.3846
g_religion	0.2	0.4472	0.8462	0.6538	0.5959	0.0	0.3846
g_religion	0.25	0.4595	0.7846	0.5769	0.5496	0.0	0.3846
g_religion	0.3	0.4766	0.7846	0.5385	0.5122	0.0	0.3846
g_religion	0.35	0.4951	0.7846	0.5	0.4743	0.0	0.3846
g_religion	0.4	0.5051	0.7692	0.4712	0.442	0.0	0.3846
g_religion	0.45	0.5275	0.7385	0.4135	0.4136	0.0003	0.3846
g_religion	0.5	0.5402	0.7231	0.3846	0.3909	0.0184	0.3846
g_religion	0.55	0.5476	0.7077	0.3654	0.3658	0.0013	0.3846
g_religion	0.6	0.561	0.7077	0.3462	0.344	0.0	0.3846
g_religion	0.65	0.5789	0.6769	0.3077	0.3207	0.0352	0.3846
g_religion	0.7	0.5789	0.6769	0.3077	0.2724	0.0	0.3846
g_religion	0.75	0.5775	0.6308	0.2885	0.2496	0.0	0.3846
g_religion	0.8	0.5652	0.6	0.2885	0.2238	0.0	0.3846
g_religion	0.85	0.5625	0.5538	0.2692	0.1964	0.0	0.3846
g_religion	0.9	0.6275	0.4923	0.1827	0.155	0.0	0.3846
g_religion	0.95	0.2917	0.1077	0.1635	0.0733	Unavailable	0.3846
g_hate	0.05	0.3923	0.8361	0.7315	0.4912	0.0	0.3609
g_hate	0.1	0.42	0.6885	0.537	0.262	0.0	0.3609
g_hate	0.15	0.6	0.6885	0.2593	0.1769	0.0	0.3609
g_hate	0.2	0.6727	0.6066	0.1667	0.0828	0.0	0.3609
g_hate	0.25	0.6889	0.5082	0.1296	0.0598	0.0	0.3609
g_hate	0.3	0.7436	0.4754	0.0926	0.0492	0.0	0.3609
g_hate	0.35	0.7632	0.4754	0.0833	0.0429	0.0	0.3609
g_hate	0.4	0.8125	0.4262	0.0556	0.0357	0.0	0.3609
g_hate	0.45	0.8	0.3934	0.0556	0.0322	0.0	0.3609
g_hate	0.5	0.8276	0.3934	0.0463	0.0308	0.0	0.3609
g_hate	0.55	0.9231	0.3934	0.0185	0.0288	0.0273	0.3609
g_hate	0.6	0.9231	0.3934	0.0185	0.0276	0.0242	0.3609
g_hate	0.65	0.92	0.377	0.0185	0.0239	0.0149	0.3609
g_hate	0.7	0.9545	0.3443	0.0093	0.0187	0.0282	0.3609
g_hate	0.75	0.9545	0.3443	0.0093	0.0158	0.0196	0.3609
g_hate	0.8	0.9524	0.3279	0.0093	0.015	0.0179	0.3609
g_hate	0.85	0.95	0.3115	0.0093	0.0115	0.0074	0.3609
g_hate	0.9	1.0	0.2787	0.0	0.0083	0.0299	0.3609
g_hate	0.95	1.0	0.2131	0.0	0.0055	0.0256	0.3609
g_morality	0.05	0.4444	0.9677	0.7009	0.7791	0.293	0.3669
g_morality	0.1	0.5556	0.9677	0.4486	0.65	0.3879	0.3669
g_morality	0.15	0.5699	0.8548	0.3738	0.5959	0.4617	0.3669
g_morality	0.2	0.6098	0.8065	0.2991	0.5528	0.5	0.3669
g_morality	0.25	0.6479	0.7419	0.2336	0.518	0.5594	0.3669
g_morality	0.3	0.6923	0.7258	0.1869	0.4858	0.5546	0.3669
g_morality	0.35	0.7258	0.7258	0.1589	0.4582	0.5279	0.3669
g_morality	0.4	0.7458	0.7097	0.1402	0.436	0.5194	0.3669
g_morality	0.45	0.7636	0.6774	0.1215	0.4136	0.5254	0.3669
g_morality	0.5	0.8	0.6452	0.0935	0.3909	0.5391	0.3669
g_morality	0.55	0.8125	0.629	0.0841	0.3707	0.526	0.3669
g_morality	0.6	0.8444	0.6129	0.0654	0.3515	0.5224	0.3669

pemra_cooccurrence.csv

Download CSV · 8 rows

pemra_cooccurrence.csv — data as published
	Religion and sect	Hate and violence	Vulgarity and morality	Women and gender	False or defamatory	State and security	Political bias	Foreign content
Religion and sect	1092	86	426	155	10	78	2	34
Hate and violence	86	162	20	15	4	5	3	0
Vulgarity and morality	426	20	1333	231	13	21	4	64
Women and gender	155	15	231	714	5	3	4	11
False or defamatory	10	4	13	5	418	15	215	1
State and security	78	5	21	3	15	141	14	3
Political bias	2	3	4	4	215	14	501	0
Foreign content	34	0	64	11	1	3	0	104

pemra_grounds_by_year.csv

Download CSV · 5 rows

pemra_grounds_by_year.csv — data as published
year	n	n_distinct	Religion and sect	Hate and violence	Vulgarity and morality	Women and gender	False or defamatory	State and security	Political bias	Foreign content
2020	1533	807	0.3477	0.0581	0.227	0.259	0.028	0.062	0.1689	0.0183
2021	1111	916	0.1728	0.0324	0.4707	0.0594	0.3141	0.0216	0.2025	0.0351
2022	191	144	0.3089	0.1414	0.6283	0.1414	0.0681	0.0628	0.0471	0.0942
2023	621	481	0.4928	0.0113	0.5395	0.3575	0.0193	0.0145	0.0097	0.0274
2024	21	20	0.0952	0.1429	0.3333	0.0952	0.0476	0.0476	0.0952	0.0952

pemra_monthly_campaigns.csv

Download CSV · 57 rows

pemra_monthly_campaigns.csv — first 50 rows; download the CSV for the full table
ym	total	campaign	unique	state	largest_campaign	largest_channel
2020-01	68	44	24	ok_v1	39	Abbtak
2020-02	60	0	60	ok_v1	0	Unavailable
2020-03	120	8	112	ok_v1	8	Neo on
2020-04	482	433	49	ok_v1	245	Hum News
2020-05	130	34	96	ok_v1	18	Samaa TV
2020-06	48	2	46	ok_v1	1	ﺍﺮﻤﻴﭘ Samaa TV ﺕﺰﻋ ﺏﺎﻨﺟ پﺁﺭﺎﺴﮐﺎﺧ ﺮﻳﺯﻭ یﺭﺩﺎﻗ ﮯﺋﺍﺮﺑ ﯽﮐ ﯽﻧﺎﺴﻟ ﻖﺷ ﻖﺷ ﻖﺷ ﻖﺷ ﻖﺷ ءﺎﻤﺳ ’’ پﺁ , ہﻄﺑﺎﺿ
2020-07	73	8	65	ok_v1	6	ARY Digital
2020-08	27	6	21	ok_v1	6	24
2020-09	423	158	265	ok_v1	84	ARY Digital
2020-10	50	5	45	ok_v1	5	GNN News
2020-11	30	0	30	ok_v1	0	Unavailable
2020-12	22	0	22	ok_v1	0	Unavailable
2021-01	115	15	100	ok_v1	7	TV ONE
2021-02	104	9	95	ok_v1	3	TV ONE
2021-03	91	21	70	ok_v1	6	ARY Digital
2021-04	44	11	33	ok_v1	11	Business Plus
2021-05	44	0	44	ok_v1	0	Unavailable
2021-06	47	0	47	ok_v1	0	Unavailable
2021-07	130	10	120	ok_v1	6	Bol Entertainment
2021-08	52	7	45	ok_v1	6	Geo News
2021-09	267	27	240	ok_v1	10	Geo News
2021-10	136	8	128	ok_v1	3	Geo News
2021-11	30	0	30	ok_v1	0	Unavailable
2021-12	51	4	47	ok_v1	4	ARY Digital
2022-01	12	0	12	ok_v1	0	Unavailable
2022-02	10	0	10	ok_v1	0	Unavailable
2022-03	31	3	28	ok_v1	2	Unavailable
2022-04	11	1	10	ok_v1	1	Unavailable
2022-05	5	0	5	ok_v1	0	Unavailable
2022-06	12	2	10	ok_v1	2	Section 3 ARY News Prime was aired at future
2022-07	13	0	13	ok_v1	0	Unavailable
2022-08	20	0	20	ok_v1	0	Unavailable
2022-09	20	0	20	ok_v1	0	Unavailable
2022-10	9	0	9	ok_v1	0	Unavailable
2022-11	3	0	3	ok_v1	0	Unavailable
2022-12	45	15	30	ok_v1	15	Section 3 ARY News Prime was aired at future
2023-01	40	11	29	ok_v1	5	Unavailable
2023-02	55	10	45	ok_v1	7	Unavailable
2023-03	96	23	73	ok_v1	7	۔ﮯﮨ ﺎﮨﺭ ﺎﺟ ﺎﻳﺍﺮﮐ
2023-04	61	13	48	ok_v1	4	Unavailable
2023-05	71	4	67	ok_v1	2	ٹﻼﭘ یﺰﮐﺮﻣ ﺎﮐ ﮯﻣﺍﺭڈ ۔ ﮯﮨ ﺎﮨﺭﺎﺟ ﺎﻴﮐ ٹﻮﻣﻭﺮﭘ
2023-06	118	28	90	ok_v1	16	Geo Ent. Nooran 20 drama world - Tere Bin, twitter below). (the sitting on go as male
2023-07	107	8	99	ok_v1	6	ٹﻼﭘ یﺰﮐﺮﻣ ﺎﮐ ﮯﻣﺍﺭڈ ۔ ﮯﮨ ﺎﮨﺭﺎﺟ ﺎﻴﮐ ٹﻮﻣﻭﺮﭘ
2023-08	49	3	46	ok_v1	2	Section 3 ARY News Prime was aired at future
2023-09	7	0	7	ok_v1	0	Unavailable
2023-10	4	0	4	ok_v1	0	Unavailable
2023-11	6	0	6	ok_v1	0	Unavailable
2023-12	7	0	7	ok_v1	0	Unavailable
2024-01	5	0	5	ok_v2	0	Unavailable
2024-02	2	0	2	ok_v2	0	Unavailable

pemra_denominators.csv

Download CSV · 3 rows

pemra_denominators.csv — data as published
denominator	n	Religion and sect (share)	Hate and violence (share)	Vulgarity and morality (share)	Women and gender (share)	False or defamatory (share)	State and security (share)	Political bias (share)	Foreign content (share)	Religion and sect (count)	Hate and violence (count)	Vulgarity and morality (count)	Women and gender (count)	False or defamatory (count)	State and security (count)	Political bias (count)	Foreign content (count)
Every filed complaint	3477	0.3141	0.0466	0.3834	0.2053	0.1202	0.0406	0.1441	0.0299	1092	162	1333	714	418	141	501	104
Each distinct text once	2342	0.2844	0.047	0.4667	0.1874	0.1516	0.0295	0.0914	0.0371	666	110	1093	439	355	69	214	87
Outside any campaign	2546	0.2863	0.0495	0.4768	0.1905	0.15	0.0275	0.0899	0.0408	729	126	1214	485	382	70	229	104

Reuse and citation

Hiba Sameen / Adaad (2026). PEMRA complaints: provisional text labels and classification examples. https://adaad.org/datasets/pakistan-pemra-complaint-classification/. Adaad’s original analysis is CC BY 4.0. Underlying third-party data retain their own terms. Please credit Adaad and the original sources.

Full source and method notes · Corrections