Journalist casualties in the Israel-Gaza war

> Around 10% of Gaza's journalists have been killed versus 2.5% of healthcare workers.

These figures are insane! The number of journalists killed crossed the threshold where Gaza journalists can be declared officially decimated, according to the historical definition of the term as 1/10 killed[1]. I noticed if we plot the journalist killings against general population killings in Gaza, the graph deviates from an approximately straight line that might be expected if journalists were merely being killed as a result of "collateral damage" amidst indiscriminated bombings, and I got nerd sniped[2] by the question: Do data give any clues Israel purposefully targets journalists, beyond the large values[3]?

After spending some days grappling with it, I think the shape of data supports it, as displayed here:

Graph: https://imgur.com/a/SWNSYOn Journalist killing data: https://pastebin.com/S1ykpSMR (from https://en.wikipedia.org/wiki/List_of_journalists_killed_in_...) General population killing data: https://pastebin.com/EGq6cgH8 (from https://www.ochaopt.org/)

Discussion:

Looking at the plot of the ratio of journalists killed by israeli forces over all killed, it appears to be slightly decreasing over time instead of keeping constant. My working hypothesis is, as journalists are dispoportionally targeted, their numbers get comparatively depleted over time, making slightly harder to kill more journalists. This can be modelled as an urn problem[4], when each ball colour has different odds of being drawn, in addition to being present in different numbers. But instead of black/white balls, we have people, and instead of being drawn out from an urn, they are drawn out of life, by being shot at[5], shelled, bombed, run over by bulldozers[6], et cetera. The urn problem with different weights on each color can be modeled by a noncentral hypergeometric distribution[7], either Wallenius' noncentral hypergeometric distribution[8], for the case where balls are sampled one by one in such a way that there is competition between the balls, or Fisher's noncentral hypergeometric distribution[9], for the case where the balls are sampled simultaneously or independently of each other.

There's some disagreement on published numbers of killed journalists. For example, the Comitee to Protect Journalists counts around 90 jornalists killed[10]. The Euro-Med Monitor over 130[11]. I'm not sure about the causes of this disagreement, but I suspect is due to who is counted as journalist, only people that actualy write reports and give face and voice to established news sources, or also their supporting crews (camera men, fixers, drivers...) and freelancers. In doubt I used data extracted from a table given by wikipedia[12], that is closer to the more conservative estimate given by CPJ and lists individual names and sources. The pattern of journalists killed each day in the dataset (1, 3, 1, 3, 1, 1, 2, 1, 1, 2, 1, 1, 1, 1, 1, 1, 5, 1, 1, 1, 2, 2, 2, 1, 1, 2, 1, 2, 6, 1, 2, 1, 2, 1, 2, 1, 4, 2, 1, 3, 1, 1, 2, 1, 1, 1, 3, 2, 1, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1) indicates they're taken down mostly one at a time, and even large individual massacres, like the recent flour bag massacre[13], kills less than 1/10000 of the total population from Gaza Strip each, so Wallenius' distribution should approximate it better. In fact, it appears it is more likely for journalists to be killed together with their families at home than together with other journalists doing fieldwork, as we can infer by reading the entries in the description field of the table. For the figures on the whole strip, I collected daily reports published in the site of United Nations Office for the Coordination of Humanitarian Affairs (OCHA), and scraped some barcharts[14].

1/3 (continues)

  import csv
  from collections import Counter
  from datetime import date, timedelta
  from itertools import accumulate, islice
  from matplotlib import pyplot as plt
  from scipy.stats import nchypergeom_wallenius
  from statistics import linear_regression
  
  def batched(iterable, n):
      # batched('ABCDEFG', 3) --> ABC DEF G
      if n 

  Python 3.11.8 (main, Feb  7 2024, 00:00:00) [GCC 13.2.1 20231011 (Red Hat 13.2.1-4)] on linux
  Type "help", "copyright", "credits" or "license()" for more information.
  
  ========== RESTART: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX ==========
  Linear regression parameters: slope = 0.00344597608865967 ; intercept = 0.0 
  
  n, odds (best Wallenius fit): (155, 75) 
  
  Top 20 (n, odds): [(155, 75), (160, 72), (145, 84), (165, 69), (170, 66), (150, 78), (175, 63), (140, 87), (135, 93), (150, 81), (185, 60), (140, 90), (180, 60), (190, 57), (195, 54), (200, 54), (135, 96), (180, 63), (205, 51), (185, 57)] 
  
  Top 20 n variation in range: 135 - 205
  Top 20 odds variation in range: 51 - 96