Independent data journalism on Pakistan. Adaad reads the numbers behind the economy, workforce and society. Every story names its sources, publishes its data and explains its limits.
In Pakistan's provincial surveys, fewer than one in five women who said they had been physically attacked reported that the police were told. Court records show how allegations can remain unresolved when accused people abscond, and how verdicts can remain contested for years.
Read the lead storyEach bar is one jurisdiction's decisions in the designated gender-based violence courts in 2023, split into the Law and Justice Commission's categories. The first two are the Commission's categories of convicted, and of acquitted after completion of trial. The rest are acquittal under sections 249-A and 265-K, which allow acquittal at any stage, proceedings stopped, proceedings initiated under section 512 while the accused is absconding, compromise and indefinite adjournment. Hover over a segment for the count.
I want a Pakistan in which belonging is secure enough to accommodate disagreement, and citizenship gives people a real claim on the state.
Read the letter →Ahmadis made up 0.07 per cent of the population recorded by religion in the 2023 census, but 26 per cent of the people the Centre for Social Justice (CSJ) recorded as accused of blasphemy between 2020 and 2025, including 141 of 362 in 2025. Most recorded accusations came from Punjab and Sindh; district totals cannot show whether accusations are followed by minority communities shrinking.
Each column is a year, split by the religion recorded for each person accused. Of the 362 people recorded in 2025, 213 were Muslim and 141 Ahmadi, against 49 Ahmadi the year before. Switch to share of the year to compare years of different size.
Across South Asia, citizens continue to protest even as democratic institutions weaken. Some movements have forced leaders from office; others have won concessions or sustained opposition under repression.
Each panel is a country, scored by V-Dem’s experts from 0, virtually no mass events, to 4, many large and small ones. Green is mobilisation for democracy and red for autocracy. Red rises above green only in Bangladesh in 2020 and the Maldives in 2013, 2015 and 2016; in Bhutan both are near zero.
In April 1952, Pakistan’s first Constituent Assembly voted 41 to 12 to postpone a decision on recognising Bengali as a state language. Ten of the 12 who opposed delay also voted for both representation across communities and protections for non-Muslim personal laws. For them, recognising more than one language was part of recognising more than one kind of Pakistani.
Each row is one of six other divisions between 1948 and 1953. Green dots are members who opposed postponing the Bengali decision in April 1952, tan those who backed it, and the bar between them is the gap. On opening East Bengal’s general seats to every community, 10 of 11 opponents of delay voted for it and none of 29 backers did.
Women’s turnout trailed men’s by about nine percentage points in 2024, close to the gap in 2018. It was nearly eighteen in Pashto-majority seats and eleven or more in the selected urban seats of most cities examined. A similar national average concealed different movements across the country.
Each dot is a National Assembly seat, placed by the share of its residents whose mother tongue is Pashto and by how far women’s turnout trailed men’s. In 2024 the gap was 17.6 points across the 37 seats over 70 per cent Pashto-speaking, against 9.0 across all seats. Switch the year or the language to compare.
Tehreek-e-Labbaik Pakistan polled 2.9mn votes in 2024 and won no National Assembly seat. In 69 seats, it polled more votes than separated the winner from the runner-up. That count shows where the party’s voters could have mattered, not where they did.
Shaded seats are the 69 where TLP finished third or lower and polled more than the winner’s margin, 61 of them in Punjab; the darker the seat, the fewer of TLP’s voters would have had to back the runner-up to change the winner. Switch views to compare how TLP’s vote and the other religious parties’ are spread.
How a computer sorts text into categories, why one label per text is often the wrong design, and what it takes to turn a classifier’s labels into a count worth publishing.
Swipe the chart to read all labels →
Each column is a month of complaints to PEMRA’s call centre, January 2020 to September 2024. Gold marks complaints in campaigns, including each campaign’s first filing: 931 of the 3,477 complaints belong to 47 campaigns of five or more near-identical complaints, most of them in a few months. Hover over a month for its largest campaign.
Two districts in Sindh: their reported forced conversions and abductions, minority populations, and figures on poverty, education, marriage, crime and the courts.
The bars show reported abduction or forced-conversion cases a year per 100,000 non-Muslim residents: 6.0 in Ghotki and 0.9 in Umerkot. The totals are 21 and 18 cases across 2021, 2023 and 2024, with 117,060 and 639,008 non-Muslim residents in Census 2023. Abduction alone is included from 2023.
“Without intervention of the public into power, without balances, without checks, there is no democracy.”
A short, thoughtful reading of the numbers shaping Pakistan. The next issue lands Wednesday 18 November 2026.