I'm starting this thread on why doing data-driven research in Pakistan is *so* hard. Here I'll document mostly my own experiences and observations. I'll try to provide comparisons when I can.

Today, I'll talk about local govt data

One pet peeve I have is that despite so many experiments with local govt in Pk, it is shocking to me that we have no systematic data on *basic* descriptives on what actually happened. Take the 2015 LG elections in KP as an example...

The province of 30million went from directly electing ~125 reps to the KP assembly to electing around 45,000 people across the province. That's a huge move towards democratization, but no one knows how many ppl ran for office, what the degree of competition was on various seats..

.. whether the various quotas for youth, non-muslims, farmers, women etc mattered for who contested and won. How can we even begin to evaluate the efficacy of this reform without knowing whether it moved the things it should have moved first?

I have tried multiple times to begin assembling KP local elections with @yasirmkh but we've repeatedly found that even the Election Commission has no idea what actually happened because no one bothered collecting data even at the district level.

We managed to assemble data on about 200 villages in a couple of districts and found some striking patterns that really should be examined systematically.

In our sample, not a single woman ran on an open seat. Many women's seats were won uncontested. The highest competition was for youth seats, not open seats. Farmer seats won with more votes than people who became nazims.

There are *huge* implications from each of these basic descriptive findings, but we won't really know if this is the case everywhere without the *data*!

It is extremely important that the govt collect these data before they rot in VC offices across the provinces.

As a counter-example, consider Nepal, a country roughly the same pop size as KP in Pakistan. Nepal held LG elections in 2017. ALL election data are available centrally with the Election Commission. -_-

And then you hear stories like this: https://twitter.com/yasirmkh/status/1278040437161254912?s=20

One day I'll do a thread on my 3-year struggle to get CSS data from FPSC.. data that are literally rotting in files in the FPSC basement, that I and my co-authors offered to digitize free of cost. Data that are easily available in India.

Since @colincookman brought this up, here's my story about political constituency shape files in Pakistan https://twitter.com/colincookman/status/1278043760337747968?s=20

Back in 2009 we were looking for shape files and the Election Commission used to have these PDFs with vector data but each constituency was separately uploaded and the underlying geo information was missing..

It seemed like the election commission in fact had the underlying shape files so we got very excited and tried to find out if we could get access to those. We discovered..

that the election commission paid a consulting firm to generate the files, but only received PDFs from the. They never asked them for or verified the underlying shapefiles, which have since been lost. ¯\\_(ツ)_/¯

In the meanwhile, *all* FIR data from India are available *online* https://www.nirvikarjassal.com/research  https://twitter.com/Sindhu_Abasi/status/1278049407653556227?s=20