National Timelines
We compute the differences between the actual and expected date counts (according to our Null Model), and create temporal heatmaps of significant time periods for each country.
Temporal focal points of selected countries based on the differences between empirical data and the Null Model. z-scores below -3 and above 3 correspond to p-values < 0.01, which means the results in all coloured cells are statistically significant. Higher z-scores (orange) correspond to positive differences between the observed and the expected date count per decade. Cells with fewer than 30 dates are masked out. Interpretation of historical events corresponding to some focal points (in green) is offered by history experts. The distributions of focal points suggests there are similarities across countries within geopolitical blocs.
There are noticeable differences in distributions of focal points (in dark orange) across countries. For Western European countries, we observe high coverage of the Medieval and Early Modern periods (until ~1800). Specific periods of interest for individual countries include, for example, the French Revolution in France (1780-90s) and the Third Reich in Germany (1930-40s).
By contrast, in East Asia the focal points are more heterogeneous. For Mongolia, the timeline focuses on the Mongolian Empire in the 13th century. Articles on Japanese and Chinese histories exhibit a strong focus on specific small time frames: the rise of the Tokugawa shogunate (1180-90s), the Kenmu Restoration (1330s) and the beginning of the Edo period (around 1600) in Japan; and the rise of the Jin (1120s), Yuan (1270s), Ming (1360s) and Qing (1640s) dynasties in China. Only with stronger European involvement in the region (starting in the mid-19th century) there is a more steady coverage.
For Central America, the timelines focus on the Age of Discovery (late 15th - early 16th centuries), and the Spanish-American Wars of Independence (first half of the 19th century).
In North America, the eras of the American Revolutionary War (end of 18th century) and the American Civil War (1860s) are most noticeable.
For different regions of Africa, historical timelines strongly focus on the periods of its occupation and colonisation (Scramble for Africa in late 19th century), and recent history following its decolonization in the 1960s.
In contrast to Southern Africa, North African national timelines focus on the Medieval history (Caliphate era), which is also the time of close interaction with Europe. The coverage seems to seize around 1300, just before the outbreak of the Black Death epidemic.
For Australia and New Zealand the peaks in 1760-80s correspond to the expeditions of James Cook discovering Oceania and South Pacific. Over the next centuries, as contacts between Europeans and the local population grew, the coverage remains stable.
Overall, the number of discovered `focal points' differs across regions. Within 30 examined Wikipedia editions, there is a disproportionate focus on histories of European countries, and the coverage of non-European states seems more intense in the periods when those states had closer interactions with Europe.
Detailed temporal focal points of all countries grouped by region:
Interpretation by History experts
Grouping countries with similar historical timelines
We represent each country as a vector of z-scores, and group together the countries whose z-score values across decades are similar both in direction and intensity. We compute pairwise cosine similarity between all countries, and apply hierarchical clustering with complete linkage on top of the obtained values.
World map of country clusters. This results from cutting the dendrogram of hierarchical clustering at a threshold t=.2. The countries within clusters have similar temporal focal points, based on the articles in 30 analysed editions, and correspond well to geopolitical regions.
We get a transnational impression on patterns of similarity among national timelines. We see, for example, that most of Africa maps to one cluster. Despite individual differences between country histories, in the analysed descriptions, history of the entire continent is reduced to the periods of its (de-)colonisation. Similarly, focal points of most of Central and South American countries are limited to the Age of Discovery and their Wars of Independence. On the contrary, Europe is separated into several clusters, as here the differences among the individual national timelines are more distinct.
This gives an impression of how the entire world groups into regions based on the extracted focal points of individual countries. Also, it illustrates that for some parts of the world (e.g. Africa and parts of Americas), the analysed timelines show a reduced view of history.
Complete dendrogram of found clusters
Focal points of national histories on the word map
Temporal focal points of all countries for the period between 1940-1949, based on the difference between the observed data and the Null Model. Colour bar shows the range of z-score values; z-scores below -3 and above 3 correspond to p-values < 0.01 (are statistically significant). Each country is coloured according to the value of its z-score in the relevant decade. Higher z-scores (orange) correspond to positive differences between the observed and the expected date count per decade. The plot covers the span of the WWII, and some of the countries with significant focus on this decade include Germany, Poland, Romania, Japan, and Indonesia (at the time Dutch East Indies occupied by the Japanese Empire).
Animated national temporal focal points on the world map between 1000 - 2015 (decade per frame)
Animated temporal focal points of all countries for the period between 1000 - 2016, based on the difference between the observed data and the Null Model. Colour bar shows the range of z-score values; z-scores below -3 and above 3 correspond to p-values < 0.01 (are statistically significant). Each country is coloured according to the value of its z-score in the relevant decade (in grey above the colour bar). Higher z-scores (orange) correspond to positive differences between the observed and the expected date count per decade.
We visualise a world map for each decade, where each country is coloured in accordance with its z-score value in the current decade. This helps to gain a transnational perspective on how the temporal foci of national historical coverage were changing during the last 1,000 years. In the plot and video below, the timelines of the countries highlighted in orange focus significantly more on the decade in question, compared to other decades and countries.
Complete dendrogram of hierarchical clustering, based on cosine similarity values for all country pairs. The clusters of similarly demonstrate that countries with similar focal points distribution largely correspond to geopolitical regions.