Where They Were Found, and Where the Reports Put Them
The case record, part 2 of 3: the places.
By Piet Speurder · · updated · 7 min read
The case conflictsDawn ParkmapplacesRhodesfield
Thirteen women, thirteen places, and eleven points where the reporting disagrees with itself. Part two from the record: the map, three suburb mix-ups, the dates that moved, and the one that moved on my own map too.
Kempton Watch, from the record as of 5 October 2026
Part one was about the number. This one is about the places.
Every woman in my register has a suburb next to her name. That is as close as the site goes, and it is a rule, not a limitation. The pin is where she was found, never where she lived or where she was last seen. The reason is simple. A pin on a house is a pin on a family, and on a street where a suspect's neighbours live. I don't know enough to put it there, and I wouldn't if I did.
The map
The map at the top of this post, twelve numbered pins on an outline of Ekurhuleni, draws the twelve women in the police count, numbered in that count's order, from Olifantsfontein in the north to Springs in the east, with one outside the metro in Moroka, Soweto. The thirteenth, the woman the police count leaves out, is on the live map on the home page, marked as outside the count. Each pin sits on the centre of its suburb. If you live in Rhodesfield or Pomona or Villa Liza, don't go looking for the spot. The spot is not on this map.
Eleven places the reports disagree
My pipeline keeps every value an outlet printed, and when two outlets, or an outlet and an official statement, say different things about the same woman, it keeps both and marks the row. On 5 October there are eleven of those: four dates, two suburbs, three ages, one name and one nationality. More than half are about where or when.
I'll take the places first, because those are the ones a map can get wrong.
Three suburbs that moved
Busisiwe Mahlangu, 15 July. The reports that give a landmark put her near the R21. Daily Maverick and Mail & Guardian say Rhodesfield. IOL's explainer says Pomona. The Inter-Ministerial Committee (IMC) list says only "Kempton Park". Rhodesfield and Pomona are both Kempton Park, on either side of the highway, five kilometres apart by centroid. The register shows Rhodesfield, because two outlets say it against one, with an asterisk, because the record doesn't pick winners quietly.
The unidentified woman in the Clayville river, 7 September. The IMC says "Clayville river in Olifantsfontain", spelling and all. IOL and joburgetc say Tembisa. The Citizen says "near Tembisa Mall". Clayville and Tembisa are next door to each other, three kilometres apart. Loose, not wrong.
Gracious Nkomo, 10 September. This is the one that matters. On the morning of 10 September an unidentified woman was recovered from a spruit in Kempton Park, near Voortrekker Road at the R21 crossing. The Citizen reported it the next day. Her family identified her a week later. The IMC list has her at number five: "Gracious Nkomo found in Kempton Park on 10 September 2026".
On 17 September, the day she was identified, a ninth body was found on Lama Street in the Villa Liza area of Dawn Park, Boksburg. That is Jabulile Ntimba, number nine on the IMC list. Gauteng Community Safety issued a statement about it the same day.
Then the round-ups started. On 21 September The Citizen wrote that "the semi-naked body of Gracious Nkomo was found on the streets of Dawn Park last week". It wasn't. Eight outlets carry 17 September or Dawn Park for her somewhere in their coverage, and a ninth source, that gov.za statement about the ninth body, got attached to her row by my own pipeline, which I come back to below. That's a 14 kilometre move and a week in the date, and it comes from two women found a week apart being folded into one paragraph.
The dates that moved
Itumeleng Kekana, 3 August, or 24 August? The Citizen reported her body found near Great North Road in Pomona on 3 August, the day it happened.
Five weeks later, in a round-up of the first four cases, the same outlet wrote: "The second victim, a 32-year-old woman, was found near the R21 on 24 August." That sentence travelled. Cape Times, Daily Maverick, IOL, Mail & Guardian, TimesLIVE and others carried 24 August from then on. The IMC list says 3 August. So does The Citizen's own first report. The register shows 3 August, with the asterisk.
Elizabeth Moselakgomo, 12 or 13 September. She was found on the Saturday evening; eNCA and SABC News reported it on the Sunday with the Sunday's date. One day, and it matters only when someone writes "found on the same day".
Vutomi Mathebula, 14 or 15 September. Found on the Monday, reported by some on the Tuesday. The Guardian and joburgetc carry the reporting date as the discovery date.
The round-up problem
I'm not picking on IOL, but their explainer "where the 9 bodies have been found" carries four values that an official statement contradicts: Pomona for Mahlangu, 24 August for Kekana, Tembisa for the Clayville body, and Dawn Park on the 17th for Nkomo. I'm not picking on The Citizen either, whose round-up contradicted The Citizen's own reporter from five weeks before. Both outlets' first reports were right, and both were there when nobody else was.
The pattern is the same every time. The first report, written by someone who was there or spoke to someone who was, is usually right. The round-up, written a week or three later from other articles, is where the mistakes live. And round-ups are what people read, because they're the ones that say "here is everything we know".
My pipeline reads round-ups too. Which brings me to my own map.
My map was wrong too
Ag. On 4 October, while merging duplicate rows, the pipeline put that gov.za sentence about the ninth body, "Another body of a woman has been discovered in Dawn Park", on Gracious Nkomo's row. Her pin moved 14 kilometres to Villa Liza. I caught it the same day and corrected it by hand, with the reason written down. Then the next morning's refresh flipped her suburb back to Villa Liza, because a gov.za sentence sits on each side and the tie broke the other way. So I pinned that too, with a reason. Both corrections are in the record's log, and her suburb now shows Rhodesfield with no dispute. What's left on her row is a date conflict, 10 or 17 September, because nine sources printed the 17th, and one of those nine is that same gov.za statement about the ninth body, which the pipeline still attaches to her. Until I detach it, her row says so, and the conflicts page shows all nine.
That's the honest version. A map built from reporting inherits the reporting's mistakes, and the only defence is to show every version and say which one you chose and why.
Ages and a name
The rest of the eleven are smaller. Dineo Motapane is 38 in five outlets, 35 in three, and 30 in one. Gracious Nkomo is 28 or 30. Elizabeth Moselakgomo is 38 everywhere except Al Jazeera, which has 20. The woman shot in Tembisa on 28 September is Boitumelo Gift Mashita in SABC News, IOL, The Citizen and two others, and Boitumelo Nhlanzi in Daily Maverick. The register shows the majority value and keeps the rest on the conflicts page.
What this means if you live there
If you're in any suburb on this map, from Olifantsfontein to Springs, the pin is your suburb, not a spot. Please don't go looking. If you know that a place in the register is wrong, and you can point me to a report that says so, write to hello@kempton.watch and I'll check it. The register shows the sources behind every pin, and the sources page says how it was built.