Information Visualization · University of Bologna · 2025

The Spatial Expansion of Chinese Higher Education

A data-driven portrait of 849 universities, from 1879 to the present

849 universities 32 provincial-level divisions (data excludes Taiwan Province and Macao SAR) 7 macro-regions Source: Wikidata 4 Research Questions

The development of China's university system has been shaped by distinct historical phases, including political transitions, institutional reforms, and economic expansion. From a handful of colonial-era institutions in the 1880s, the system grew into a network spanning every corner of a continent-sized country. This project uses Wikidata records to trace four dimensions of that story: temporal rhythm, provincial inequality, regional sequencing, and geographic drift.

849
unique universities
after deduplication
1879
earliest institution
in dataset
1958
peak founding year
(61 universities)
63%
have valid
GPS coordinates
1

When did China build its universities?

Annual count of newly founded universities (bars) overlaid with cumulative total (line). Four political inflection points are annotated. The x-axis begins at 1895 — only 3 institutions predate this.

· · ·
Key finding: University founding in China is not gradual — it arrives in four distinct surges, each triggered by a political event. The 1950s alone account for 213 new institutions (25% of the entire dataset), dwarfing every other period.
Why these surges? Each wave reflects a distinct political logic: 1952 was about socialist transformation of education; 1958 about mass mobilization under the Great Leap Forward; 1978 about recouping lost human capital after a decade of closures; and 1999 about absorbing labor-market pressure from a growing youth population while stimulating domestic demand through education spending.

The surges above were not distributed evenly across the country. But simply counting universities hides a deeper question — are the biggest provinces really the "densest" in higher education?

2

Counting universities — or counting per capita? Two maps of Chinese higher education

Left: a bubble map showing absolute university counts per province (bubble size proportional to number of universities). Right: a ranked bar chart — toggle between absolute count and universities per million inhabitants to see how the ranking shifts when province size is taken into account.

· · ·
East North Central South Northeast Southwest Northwest

👇 Click a button — watch the ranking rearrange

Key finding — normalization reshuffles the story: By absolute count, Jiangsu (50), Guangdong (48), Beijing (40), and Liaoning (38) lead. But once population is factored in, the ranking changes dramatically: Beijing (1.83 per million) and Hong Kong (1.62) dominate by a wide margin, while Guangdong — second in absolute terms — drops to the 17th position per capita (0.38). Perhaps most striking, Shaanxi climbs to 4th place per capita (0.78), challenging the simple "coastal vs. interior" narrative.
What this reveals: The absolute view captures where universities are; the per-capita view captures where higher education is accessible. Guangdong and Shandong host many universities but also enormous populations (126M and 102M respectively), diluting density. Beijing and Hong Kong show the imprint of political and financial centrality — elite institutions deliberately concentrated in administrative hubs. Shaanxi's high density reflects Xi'an's historical role as a Cold War-era relocation destination for coastal universities (see RQ3). Chongqing's low ranking (0.09 per million) is an artifact of Wikidata coverage rather than reality — a reminder that data completeness varies by province. The per-capita ranking inherits this uneven coverage: provinces with denser Wikidata records appear inflated relative to those without, so treat it as directional rather than precise.

Whether measured in absolute counts or per capita, today's distribution is a snapshot. But how did it come to be? The next chart reveals when each region's stock of universities was actually built — and why the per-capita ranking looks the way it does.

3

When did the imbalance start? A regional timeline

If the East has more universities today, did it always? This heatmap reveals when each region's stock of universities was actually built. Color intensity shows what share of a region's total universities were founded in each decade.

▶ Pattern: Coastal regions show a slow, organic trickle over 100 years. ▶ Pattern: Interior regions show a sudden, massive spike concentrated almost entirely in the 1950s.

· · ·
Key finding: The normalized view reveals a clear coastal-to-interior gradient. The East shows sustained activity from the 1900s, with no single decade exceeding 20% — a sign of organic, diversified growth. In contrast, the Northwest and Southwest are almost entirely blank until 1950, when a single decade accounts for over 40% of their total — a signature of top-down, planned-economy redistribution.
Regional patterns: The Northeast's early activity (1940s) reflects Japanese colonial infrastructure. Central China's 1950s surge was partly driven by the deliberate relocation of coastal institutions inland — a Cold War strategy to disperse assets away from the coastline. The Southwest only approached parity after the xibu da kaifa (Western Development) campaign in the 2000s. The consistent early shading in East and South, decade after decade, illustrates that regional inequality in higher education was durable — and largely structural.

Connecting back to RQ2: This temporal view also explains the per-capita ranking. Shaanxi's high density (0.78 per million) isn't a natural endowment — it's the direct legacy of the 1950s relocations, when Xi'an became a designated inland hub. Beijing's dominance (1.83 per million) reflects a century of deliberate concentration of elite national institutions in the political capital. And the Southwest's persistently low per-capita numbers are the downstream consequence of having virtually zero higher education infrastructure before 1950 — a sprint from a standing start that has yet to close the gap.

If the balance of university founding shifted over time, did the geographic center of gravity shift too? The final map traces this movement across 150 years.

4

Did China's educational center of gravity shift?

Mean geographic center of universities founded in each decade, plotted as a trajectory on an interactive map. Circle size reflects the number of universities with valid GPS coordinates in that decade. The trajectory is confined to a surprisingly small region — the middle and lower Yangtze basin.

· · ·

Circle size = universities with valid coordinates · Color gradient: blue (1890s) → red (2020s) · Dotted line = trajectory · Scroll to zoom

Key finding: The mean center sits in the central plains throughout the 150-year span. It drifted modestly northward through the Republican period (1890s–1940s), then moved steadily southward from the 1950s through the 2000s — a 50-year drift of roughly 5° latitude. From 2010 onward, the center shifts noticeably eastward, suggesting a partial reversal toward the Yangtze Delta.
Why these movements? The 1940s northward jump reflects a surge of foundings in occupied Northeast China (Liaoning, Jilin, Heilongjiang) and Beijing during this turbulent decade — wartime relocations of coastal universities to Sichuan and Yunnan went the other direction (south), but their effect is masked because the Northeast cluster dominates the mean. The sustained southward shift after 1950 was driven by forced inland redistribution of Shanghai and Nanjing institutions, new construction in Yunnan, Guizhou, and Guangxi, and the overall 1950s expansion into lower-latitude provinces. The post-2010 eastward rebound likely reflects newer private and applied universities in Zhejiang and Jiangsu. Caution: this trajectory is based on the 63% of universities with valid GPS coordinates and under-weights provinces with poor coverage (notably Inner Mongolia, 11%) — the true center likely sits further north than shown.

Limitations & Data Quality

Six known issues that affect interpretation of the findings above

Coverage Gap

849 records represent roughly 65% of China's ~1,300 bachelor-degree-granting institutions. The dataset is biased toward historically prominent or internationally visible universities, underrepresenting newer regional colleges and vocational schools.

Missing Coordinates

Only ~63% of records include valid GPS coordinates. Inner Mongolia (11%), Shanxi (43%), and Henan (50%) are most affected, introducing systematic bias into the RQ4 mean-center trajectory.

Founding Year Gaps

Some records lack an inceptionYear; others record reorganization dates rather than true founding years. Post-2010 data is sparse — recent institutions are likely under-documented in Wikidata.

Scope & Normalization

Analysis is restricted to mainland China and Hong Kong SAR. Taiwan is excluded on historical grounds (independent system post-1949); Macau is excluded because Wikidata returns no universities for it. Per-capita figures in RQ2 use 2020 census population, so they reflect current density — not historical density at the time each institution was founded.

Definition Scope

The SPARQL query targets Wikidata class Q3918 ("university"), which includes military academies, party schools, and research institutes that may not qualify as conventional higher education institutions.

Temporal Recency

Data reflects Wikidata's state at query time (2025). The apparent drop-off after 2010 should not be interpreted as a real slowdown — it reflects documentation lag, not declining institutional growth.