General Remarks

  • The e4D interface has not yet been optimised for Internet Explorer. Please use Firefox, Chrome, or Safari instead. IE support is on its way, however. We face some problems with the new Safari 5.1 (OSX Lion only), too.
  • The performance of the interface depends on two aspects: the map-rendering server at the Leipzig Computing Centre and the JavaScript performance of your browser. In case the interface feels sluggish, try another map first – for example: Google Physical. If this does not help, update your browser software to the newest release: JavaScript performance has been hugely improved over the last couple of months for Safari, Chrome, and Firefox.
  • Datasets with less than 1,000 single points should render immediately. Sets with 10,000 items may take some time, depending on your machine and browser. If the browser sends a timeout message, simplyignore it: up to 40 seconds for 10,000 items sounds reasonable. You can even try more than 10,000 items, but 20,000 should be the limit. e4D has not been designed for browsing your whole repository at once.
  • Always keep in mind: e4D makes complex data accessible in an easy way. But aggregation always means reduction of complexity. We rely on the quality of the datasets underneath – the data may have errors, or may have been processed in odd ways. Hence we would encourage you to take a close look at the results tables in the lower part of the interface to get an idea of the underlying data.

europeana4D in less than 10 minutes


The central component of Europeana.4D is the map. The current prototype features several well-known concepts in terms of usability. Generally, it works like better-known map services on the web, such as Google Maps. It supports zooming as well as drag and drop for navigation. Additionally, it is possible to display historic maps for better contextualisation of the displayed items.

historical map of 1783 - showing publications of Shakespeare


The map widget consists of a geographical map, which is overlaid with bubble glyphs. The glyphs are circles, with each of the glyphs represents one or more items of the dataset. We chose a circle representation on the map to display data densities instead of, for instance, a heat map, in order to allow multiple query results to be shown together, and to make every data item individually accessible through its graphical representation.
In addition, to avoid visual clutter, we require that the glyphs not overlap each other. To achieve this goal, we merge circles based on their size, distances, and the current zoom level in an iterative process.

Historical maps

For datasets with historical content we provide a great variety of historical maps, which show the political situation in specific eras. We offer 23 different historical maps from 2000 BC to 1994 AD provided by Thinkquest. Imagine a dataset that contains elements with time stamps that fall within in the Middle Ages alone. Unfortunately, one map cannot represent all political situations in the time span “Middle Ages,” and a “mean” between country borders would not only be hard to compute, it would also misrepresent the facts. For that reason, we choose the map for the median time stamp of all data objects of the dataset as default. For instance, this could be the historical map of 800. We also allow the user to switch maps.

With this feature, the representation of the elements on the map that shows the political situation of that time can be more informative than visualising the results on a contemporary political map.

Dynamic aggregation of data

To avoid a confusingly large number of dots for large results or datasets, the interface automatically aggregates points depending on the zoom level. To ensure non-overlapping circles, we perform a variant of the agglomerative clustering algorithm.

The underlying Wikileaks Afghanistan war logs dataset shows key incidents in the Afghanistan conflict. (a) Each circle with minimum radius rmin represents one incident. The high degree of overplotting makes it hard to access each circle individually, and to determine the conflict centres. (b) 307 elements are merged to 52 non-overlapping circles. The composition of circles reflects that there were a lot of incidents in the south. The bigger circle in the northeast shows that there have been a lot of key incidents in the capital Kabul.

Placename tag cloud

Each of the map’s circles is associated with a tag cloud of the most frequent place names, including their quantity and whether they were provided by the given data. The size of the literals of a place name is proportional to its quantity in the corresponding circle.

This feature enables on-demand labelling of points and also provides a preview of how a glyph arising from agglomeration would split if zoomed in. If the data offers different levels of detail for a place, we choose the label dependent on the current zoom level. We distinguish four semantic levels: country, region (which can be, for instance, a state or a countryside), city, and borough (which can be a district, specific place or an address of the given city).

Time line

Since the timeline is a rarely used control in web applications, it wasn’t possible to reuse existing GUI concepts. The timeline is primarily used for drill-downs in time. Another functionality is the display of individual datasets for a given point in time. This is done via a simple mouse over.



Fixed time selection

The time widget allows both the clicking on one bin and the selection of a time range using a mouse drag gesture. A toolbar is then shown that offers the possibility of modifying the left or right border of the selected time era and adjusting a feather range beyond the selection borders to smooth the transitions between selected and non-selected elements.

This triggers a weighted colouring of the non-selected circles inside the map. Additionally, a user can display time-dependent connections between data items. For each bin within the selection a minimum spanning tree between the corresponding circles’ centres is displayed on the map. This helps to detect geo-spatial dependencies in short time periods. Finally, the user can drag the selected time era manually or by animation. We then update the circles on-demand to reflect how locations change over time. A play button starts the animation and the selected time span moves smoothly over the entire time range. The advantage of this feature is that the user can direct his attention to the changes inside the map. Finally, a fixed selection can be turned into a refined query.


The third method for displaying the result set is a table. The table also interacts with the timeline. If a duration is selected, the items will be highlighted.

The boxes of the table widget can also be changed by selecting a map or timeplot. A temporary selection causes highlighting of a box’s border and a fixed selection fills the box’s area. Additionally, the table can be used for the selection and the de-selection of single elements. Via this feature, the user is able to adapt a selection by adding or deleting specific elements. Finally, the user can switch the display mode of the table, so that only selected items are shown, which simplifies the browsing through the final selection.

Selection tools

Several selection tools on the map and on the timeline allow drill-down in time and space. The map features selectors for polygons, circles, and countries. These also work on the historic maps.

The timeline allows selections of points in time and durations. Durations can be fuzzy. It’s also possible to drag a selection on the timeline, during which the dots on the map change in real time.

Concurrent datasets

An advanced feature of this site is concurrent searching. Using this mode allows combining up to four (limited for clearness) searches. The user has the full possibilities given for a single search. This includes selection editing and display of result sets.



Intra dataset comparison

One of the major features of the detail widget is the ability to export elements of a fixed selection from a dataset as a new individual input set.

Hence, the temporal comparison of different geographical regions from one dataset is feasible as well as the geographical comparison of different time periods.


Another advanced feature is the possibility to connect the dots by time. This mode allows one to follow persons or ideas of European cultural heritage in time. By dragging the selected duration the connections can be adjusted.