Sandık Atlası

Sandık AtlasıAnalysis of the Week › Change in İstanbul
Analysis of the Week

Where did the votes shift in İstanbul? Neighborhood by neighborhood, 2018 → 2023

Between two parliamentary elections, İstanbul's overall picture seems to have changed little. But descend to the neighborhood level and the movement is far sharper: some districts swung one way while the one right next door behaved completely differently.

16 June 2026 · Official YSK data · neighborhood level · parliamentary 2018 → 2023

Method and stance. This analysis contains no commentary; it shows the change in each party's vote share in a neighborhood between the two elections. The comparison is İstanbul, the last two parliamentary elections (2018 → 2023). Only neighborhoods with at least 500 voters were included (in small units, a swing of a few votes distorts the percentage). Neighborhoods whose population changed substantially were excluded — those whose voter count changed by more than 15% are omitted (the shift there may not be a change of mind but a new population that moved in). Institutional units such as prisons were not counted as neighborhoods. A total of 647 neighborhoods. No candidate/person is compared; everything is at the party level.

First, the big picture

Across İstanbul as a whole, the parliamentary vote changed from 2018 to 2023 as follows (as a percentage of valid votes):

İstanbul overall — change in party share (2018 → 2023)

Parliamentary election, as a percentage of valid votes. Right = rose, left = fell.

percentage-point change · left = fell, right = rose AK Parti −7.2 42.6% → 35.4% Yeşil Sol / HDP −4.4 12.5% → 8.1% MHP −2.4 8.4% → 6.0% İYİ Parti ±0.1 8.1% → 8.2% CHP +1.8 26.5% → 28.3%
AK Parti was the party that fell the most, at −7.2 points. CHP's rise (+1.8) is modest in the aggregate — meaning most of the vote vacated by AK Parti did not go to a single party but was spread among the new parties that entered the picture in 2023 (e.g. Yeniden Refah, Zafer, TİP, Memleket).

The aggregate picture seems to say "little changed." The real story is in the neighborhoods.

The most striking finding: two opposite movements at once in central districts

In İstanbul's central, cosmopolitan districts (Cihangir, Beyoğlu, Şişli, Adalar), two things happened together in a single picture: Yeşil Sol / HDP's share fell sharply, CHP's share rose — and that, too, in the same neighborhoods.

Opposite directions in the same neighborhood: Yeşil Sol/HDP ↓   CHP ↑

Selected central neighborhoods · vote-share change (points), 2018 → 2023. Left (purple) = Yeşil Sol/HDP decline, right (red) = CHP rise.

0 · scale: 5 points ≈ 32px ← Yeşil Sol/HDP fell CHP rose → Şişli · Bozkurt −28.8 +12.1 Adalar · Kınalıada −25.4 +15.3 Beyoğlu · Kılıçali Paşa −23.4 +12.5 Beyoğlu · Ömer Avni −23.4 +10.5
In the same district, purple stretches left and red stretches right: where Yeşil Sol/HDP's share melted away, CHP's share grew.

This is the overlap of two pictures in the same place: the neighborhoods where Yeşil Sol/HDP fell the most are, to a large degree, the neighborhoods where CHP rose the most. The fact that the two moved in opposite directions suggests that in these cosmopolitan districts the left/opposition vote concentrated under the CHP umbrella.

Caution — what this does and does not prove. Ballot-box data does not show who voted for whom. We cannot say "Yeşil Sol voters switched to CHP"; we only see that two parties' shares changed in opposite directions in the same neighborhood. Individual-level transfer cannot be inferred with certainty from aggregate data (the limit of ecological inference). The numbers are an indicator; not proof.

AK Parti: no neighborhood of growth in İstanbul

Looking across all 647 neighborhoods, what stands out is this: there is not a single İstanbul neighborhood where AK Parti meaningfully increased its share. Even in the neighborhood where it "rose" the most, the increase is below 1% (Silivri/Akören, +0.6 points); everywhere else it was either flat or falling. And where it fell, it fell most in its own peripheral stronghold neighborhoods — the northern villages of Beykoz, Çatalca, Tuzla — and markedly in some urban neighborhoods:

District · Neighborhood20182023Change
Beykoz · Poyrazköy74.6%57.0%−17.7
Kadıköy · Dumlupınar41.4%24.4%−17.0
Çatalca · Karacaköy Merkez52.9%37.8%−15.2
Beykoz · Cumhuriyetköy60.5%45.3%−15.1
Tuzla · Orhanlı59.8%45.0%−14.8
Sultangazi · Malkoçoğlu72.7%59.7%−13.1

Kadıköy/Dumlupınar is a perfect mirror: AK Parti lost −17 points here while CHP gained +15 points.

CHP: rose most in central and higher-income districts

The neighborhoods where CHP increased its share the most are geographically consistent: the city center and relatively upper-income districts.

District · Neighborhood20182023Change
Adalar · Kınalıada33.0%48.4%+15.3
Kadıköy · Dumlupınar27.5%42.3%+14.9
Beyoğlu · Kılıçali Paşa41.8%54.3%+12.5
Şişli · Bozkurt41.1%53.1%+12.1
Üsküdar · Kandilli50.3%61.9%+11.6
Beşiktaş · Akat54.3%64.5%+10.1

Where CHP fell, by contrast, was İstanbul's western countryside: the villages of Silivri and Çatalca (between −6 and −11 points).

İYİ Parti: flat and scattered

İYİ Parti barely moved across İstanbul as a whole (8.1% → 8.2%). At the neighborhood level, too, the movements were small and scattered: a limited rise in a few western/upper-income neighborhoods (the most being Çatalca/Fatih +8.2), and a decline in rural Çatalca. It formed no clear geographic pattern.


Data and method. Source: Supreme Election Council (acikveri.ysk.gov.tr), neighborhood level. Elections: the 24 June 2018 and 14 May 2023 parliamentary elections. Metric: the percentage-point change in each party's valid vote share in a neighborhood between the two elections. Neighborhoods included: those with ≥500 voters in both elections and a voter-count change below 15%; institutional units such as prisons excluded (647 neighborhoods). Alliance total rows were excluded to avoid double counting. Yeşil Sol/HDP were combined as a single lineage across 2018 (HDP) and 2023 (Yeşil Sol Parti). Individual vote transfer cannot be inferred from aggregate data.

← Back to Analysis of the Week · Sandık Atlası